WO2025007355A1 - 编解码方法、码流、编码器、解码器以及存储介质 - Google Patents

编解码方法、码流、编码器、解码器以及存储介质 Download PDF

Info

Publication number
WO2025007355A1
WO2025007355A1 PCT/CN2023/106178 CN2023106178W WO2025007355A1 WO 2025007355 A1 WO2025007355 A1 WO 2025007355A1 CN 2023106178 W CN2023106178 W CN 2023106178W WO 2025007355 A1 WO2025007355 A1 WO 2025007355A1
Authority
WO
WIPO (PCT)
Prior art keywords
node
nodes
current
attribute
value
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Ceased
Application number
PCT/CN2023/106178
Other languages
English (en)
French (fr)
Other versions
WO2025007355A9 (zh
Inventor
孙泽星
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Guangdong Oppo Mobile Telecommunications Corp Ltd
Original Assignee
Guangdong Oppo Mobile Telecommunications Corp Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Guangdong Oppo Mobile Telecommunications Corp Ltd filed Critical Guangdong Oppo Mobile Telecommunications Corp Ltd
Priority to PCT/CN2023/106178 priority Critical patent/WO2025007355A1/zh
Priority to CN202380098588.6A priority patent/CN121241567A/zh
Publication of WO2025007355A1 publication Critical patent/WO2025007355A1/zh
Publication of WO2025007355A9 publication Critical patent/WO2025007355A9/zh
Priority to US19/423,914 priority patent/US20260113485A1/en
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/60—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using transform coding
    • H04N19/61—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using transform coding in combination with predictive coding
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
    • H04N19/102—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or selection affected or controlled by the adaptive coding
    • H04N19/103—Selection of coding mode or of prediction mode
    • H04N19/105—Selection of the reference unit for prediction within a chosen coding or prediction mode, e.g. adaptive choice of position and number of pixels used for prediction
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
    • H04N19/134—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or criterion affecting or controlling the adaptive coding
    • H04N19/167—Position within a video image, e.g. region of interest [ROI]
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
    • H04N19/169—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding
    • H04N19/187—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding the unit being a scalable video layer
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
    • H04N19/189—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the adaptation method, adaptation tool or adaptation type used for the adaptive coding
    • H04N19/196—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the adaptation method, adaptation tool or adaptation type used for the adaptive coding being specially adapted for the computation of encoding parameters, e.g. by averaging previously computed encoding parameters
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/30—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using hierarchical techniques, e.g. scalability
    • H04N19/33—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using hierarchical techniques, e.g. scalability in the spatial domain
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/50—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding
    • H04N19/597—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding specially adapted for multi-view video sequence encoding

Definitions

  • the embodiments of the present application relate to the field of point cloud encoding and decoding technology, and in particular, to an encoding and decoding method, a bit stream, an encoder, a decoder, and a storage medium.
  • G-PCC geometry-based point cloud compression
  • the geometry information and attribute information of the point cloud are encoded separately.
  • the attribute encoding of G-PCC can include: Predicting Transform (PT), Lifting Transform (LT) and Region Adaptive Hierarchical Transform (RAHT).
  • PT Predicting Transform
  • LT Lifting Transform
  • RAHT Region Adaptive Hierarchical Transform
  • the nodes of each layer need to be transformed, predicted, encoded and decoded in turn, which will increase the complexity of RAHT attribute transformation encoding and decoding, and thus fail to effectively remove the redundancy of attributes, resulting in low attribute coding efficiency.
  • the embodiments of the present application provide a coding and decoding method, a bit stream, an encoder, a decoder and a storage medium, which can reduce the time complexity of attribute coding and decoding, improve the attribute coding and decoding efficiency of point clouds, and thus improve the coding and decoding performance of point clouds.
  • an embodiment of the present application provides a decoding method, which is applied to a decoder, and the method includes:
  • the attribute reconstruction value of the voxel node of the current unit is determined according to the first quantity and the second quantity.
  • the second processor is used to execute the method described in the first aspect when running a computer program.
  • FIG2B is a schematic diagram of a data storage format corresponding to a point cloud image
  • FIG3 is a schematic diagram of a network architecture for point cloud encoding and decoding
  • FIG5B is a schematic diagram of a high plane position in the Z-axis direction
  • FIG. 7A is a schematic diagram of a plane identification information
  • FIG7B is a schematic diagram of another type of planar identification information
  • FIG8 is a schematic diagram of sibling nodes of a current node
  • FIG9 is a schematic diagram of the intersection of a laser radar and a node
  • FIG10 is a schematic diagram of neighborhood nodes at the same partition depth and the same coordinates
  • FIG11 is a schematic diagram of a current node being located at a low plane position of a parent node
  • FIG12 is a schematic diagram of a high plane position of a current node located at a parent node
  • FIG13 is a schematic diagram of predictive coding of planar position information of a laser radar point cloud
  • FIG14 is a schematic diagram of IDCM encoding
  • FIG15 is a schematic diagram of coordinate transformation of a rotating laser radar to obtain a point cloud
  • FIG16 is a schematic diagram of predictive coding in the X-axis or Y-axis direction
  • FIG17A is a schematic diagram showing an angle of the Y plane predicted by a horizontal azimuth angle
  • FIG17B is a schematic diagram showing an angle of predicting the X-plane by using a horizontal azimuth angle
  • FIG18 is another schematic diagram of predictive coding in the X-axis or Y-axis direction
  • FIG19A is a schematic diagram of three intersection points included in a sub-block
  • FIG19B is a schematic diagram of a triangular facet set fitted using three intersection points
  • FIG19C is a schematic diagram of upsampling of a triangular face set
  • FIG20 is a schematic diagram of a distance-based LOD construction process
  • FIG21 is a schematic diagram of a visualization result of a LOD generation process
  • FIG22 is a schematic diagram of an encoding process for attribute prediction
  • FIG. 23 is a schematic diagram of the composition of a pyramid structure
  • FIG. 24 is a schematic diagram showing the composition of another pyramid structure
  • FIG25 is a schematic diagram of an LOD structure for inter-layer nearest neighbor search
  • FIG26 is a schematic diagram of a nearest neighbor search structure based on spatial relationship
  • FIG27A is a schematic diagram of a coplanar spatial relationship
  • FIG27B is a schematic diagram of a coplanar and colinear spatial relationship
  • FIG27C is a schematic diagram of a spatial relationship of coplanarity, colinearity and copointness
  • FIG30 is a schematic diagram of intra-layer prediction based on fast search
  • FIG31 is a schematic diagram of a block-based neighborhood search structure
  • FIG32 is a schematic diagram of a coding process of a lifting transformation
  • FIG33 is a schematic diagram of a RAHT transformation structure
  • FIG34 is a schematic diagram of a RAHT transformation process along the x, y, and z directions;
  • FIG35A is a schematic diagram of a RAHT forward transformation process
  • FIG35B is a schematic diagram of a RAHT inverse transformation process
  • FIG36 is a schematic diagram of the structure of an attribute coding block
  • FIG37 is a schematic diagram of the overall process of RAHT attribute prediction transform coding
  • FIG38 is a schematic diagram of a neighborhood prediction relationship of a current block
  • FIG39 is a schematic diagram of a process for calculating attribute transformation coefficients
  • FIG40 is a schematic diagram of the structure of a RAHT attribute inter-frame prediction coding
  • FIG41 is a flowchart diagram 1 of a decoding method provided in an embodiment of the present application.
  • FIG42 is a schematic diagram of the structure of a RAHT attribute coding layer provided in an embodiment of the present application.
  • FIG43 is a second flowchart of a decoding method provided in an embodiment of the present application.
  • FIG44 is a flowchart diagram 1 of an encoding method provided in an embodiment of the present application.
  • FIG45 is a second flow chart of an encoding method provided in an embodiment of the present application.
  • FIG46 is a third flow chart of an encoding method provided in an embodiment of the present application.
  • FIG47 is a schematic diagram of the composition structure of an encoder provided in an embodiment of the present application.
  • FIG48 is a schematic diagram of a specific hardware structure of an encoder provided in an embodiment of the present application.
  • FIG49 is a schematic diagram of the composition structure of a decoder provided in an embodiment of the present application.
  • FIG50 is a schematic diagram of a specific hardware structure of a decoder provided in an embodiment of the present application.
  • Figure 51 is a schematic diagram of the composition structure of a coding and decoding system provided in an embodiment of the present application.
  • first ⁇ second ⁇ third involved in the embodiments of the present application are only used to distinguish similar objects and do not represent a specific ordering of the objects. It can be understood that “first ⁇ second ⁇ third” can be interchanged in a specific order or sequence where permitted, so that the embodiments of the present application described here can be implemented in an order other than that illustrated or described here.
  • Point Cloud is a three-dimensional representation of the surface of an object.
  • Point cloud (data) on the surface of an object can be collected through acquisition equipment such as photoelectric radar, lidar, laser scanner, and multi-view camera.
  • a point cloud is a set of irregularly distributed discrete points in space that express the spatial structure and surface properties of a three-dimensional object or scene.
  • FIG1A shows a three-dimensional point cloud image
  • FIG1B shows a partial magnified view of the three-dimensional point cloud image. It can be seen that the point cloud surface is composed of densely distributed points.
  • Two-dimensional images have information expressed at each pixel point, and the distribution is regular, so there is no need to record its position information additionally; however, the distribution of points in point clouds in three-dimensional space is random and irregular, so it is necessary to record the position of each point in space in order to fully express a point cloud.
  • each position in the acquisition process has corresponding attribute information, usually RGB color values, and the color value reflects the color of the object; for point clouds, in addition to color information, the attribute information corresponding to each point is also commonly the reflectance value, which reflects the surface material of the object. Therefore, point cloud data usually includes the position information of the point and the attribute information of the point. Among them, the position information of the point can also be called the geometric information of the point.
  • the geometric information of the point can be the three-dimensional coordinate information of the point (x, y, z).
  • the attribute information of the point can include color information and/or reflectivity, etc.
  • reflectivity can be one-dimensional reflectivity information (r); color information can be information on any color space, or color information can also be three-dimensional color information, such as RGB information.
  • R represents red (Red, R)
  • G represents green (Green, G)
  • B blue (Blue, B).
  • the color information may be luminance and chrominance (YCbCr, YUV) information, where Y represents brightness (Luma), Cb (U) represents blue color difference, and Cr (V) represents red color difference.
  • the points in the point cloud may include the three-dimensional coordinate information of the points and the reflectivity value of the points.
  • the points in the point cloud may include the three-dimensional coordinate information of the points and the three-dimensional color information of the points.
  • a point cloud obtained by combining the principles of laser measurement and photogrammetry may include the three-dimensional coordinate information of the points, the reflectivity value of the points and the three-dimensional color information of the points.
  • Figures 2A and 2B show a point cloud image and its corresponding data storage format.
  • Figure 2A provides six viewing angles of the point cloud image
  • Figure 2B consists of a file header information part and a data part.
  • the header information includes the data format, data representation type, the total number of point cloud points, and the content represented by the point cloud.
  • the point cloud is in the ".ply" format, represented by ASCII code, with a total number of 207242 points, each Each point has three-dimensional coordinate information (x, y, z) and three-dimensional color information (r, g, b).
  • Point clouds can be divided into the following categories according to the way they are obtained:
  • Static point cloud the object is stationary, and the device that obtains the point cloud is also stationary;
  • Dynamic point cloud The object is moving, but the device that obtains the point cloud is stationary;
  • Dynamic point cloud acquisition The device used to acquire the point cloud is in motion.
  • point clouds can be divided into two categories according to their usage:
  • Category 1 Machine perception point cloud, which can be used in autonomous navigation systems, real-time inspection systems, geographic information systems, visual sorting robots, disaster relief robots, etc.
  • Category 2 Point cloud perceived by the human eye, which can be used in point cloud application scenarios such as digital cultural heritage, free viewpoint broadcasting, 3D immersive communication, and 3D immersive interaction.
  • Point clouds can flexibly and conveniently express the spatial structure and surface properties of three-dimensional objects or scenes. Point clouds are obtained by directly sampling real objects, so they can provide a strong sense of reality while ensuring accuracy. Therefore, they are widely used, including virtual reality games, computer-aided design, geographic information systems, automatic navigation systems, digital cultural heritage, free viewpoint broadcasting, three-dimensional immersive remote presentation, and three-dimensional reconstruction of biological tissues and organs.
  • Point clouds can be collected mainly through the following methods: computer generation, 3D laser scanning, 3D photogrammetry, etc.
  • Computers can generate point clouds of virtual three-dimensional objects and scenes; 3D laser scanning can obtain point clouds of static real-world three-dimensional objects or scenes, and can obtain millions of point clouds per second; 3D photogrammetry can obtain point clouds of dynamic real-world three-dimensional objects or scenes, and can obtain tens of millions of point clouds per second.
  • 3D photogrammetry can obtain point clouds of dynamic real-world three-dimensional objects or scenes, and can obtain tens of millions of point clouds per second.
  • the number of points in each point cloud frame is 700,000, and each point has coordinate information xyz (float) and color information RGB (uchar).
  • point cloud compression has become a key issue in promoting the development of the point cloud industry.
  • the point cloud is a collection of massive points, storing the point cloud will not only consume a lot of memory, but also is not conducive to transmission. There is also not enough bandwidth to support direct transmission of the point cloud at the network layer without compression. Therefore, the point cloud needs to be compressed.
  • the point cloud coding framework that can compress point clouds can be the geometry-based point cloud compression (G-PCC) codec framework or the video-based point cloud compression (V-PCC) codec framework provided by the Moving Picture Experts Group (MPEG), or the AVS-PCC codec framework provided by AVS.
  • the G-PCC codec framework can be used to compress the first type of static point cloud and the third type of dynamically acquired point cloud, which can be based on the point cloud compression test platform (Test Model Compression 13, TMC13), and the V-PCC codec framework can be used to compress the second type of dynamic point cloud, which can be based on the point cloud compression test platform (Test Model Compression 2, TMC2). Therefore, the G-PCC codec framework is also called the point cloud codec TMC13, and the V-PCC codec framework is also called the point cloud codec TMC2.
  • FIG3 is a schematic diagram of a network architecture of a point cloud encoding and decoding provided by the embodiment of the present application.
  • the network architecture includes one or more electronic devices 13 to 1N and a communication network 01, wherein the electronic devices 13 to 1N can perform video interaction through the communication network 01.
  • the electronic device can be various types of devices with point cloud encoding and decoding functions.
  • the electronic device can include a mobile phone, a tablet computer, a personal computer, a personal digital assistant, a navigator, a digital phone, a video phone, a television, a sensor device, a server, etc., which is not limited by the embodiment of the present application.
  • the decoder or encoder in the embodiment of the present application can be the above-mentioned electronic device.
  • the electronic device in the embodiment of the present application has a point cloud encoding and decoding function, generally including a point cloud encoder (ie, encoder) and a point cloud decoder (ie, decoder).
  • a point cloud encoder ie, encoder
  • a point cloud decoder ie, decoder
  • the point cloud data is first divided into multiple slices by slice division.
  • the geometric information of the point cloud and the attribute information corresponding to each point are encoded separately.
  • FIG4A shows a schematic diagram of the composition framework of a G-PCC encoder.
  • the geometric information is transformed so that all point clouds are contained in a bounding box, and then quantized.
  • This step of quantization mainly plays a role in scaling. Due to the quantization rounding, the geometric information of a part of the point cloud is the same, so whether to remove duplicate points is determined based on parameters.
  • the process of quantization and removal of duplicate points is also called voxelization. Then the bounding box is divided into an octree or a prediction tree.
  • arithmetic coding is performed on the points in the divided leaf nodes to generate a binary geometric bit stream; or, arithmetic coding is performed on the intersection points (Vertex) generated by the division (surface fitting is performed based on the intersection points) to generate a binary geometric bit stream.
  • attribute coding After the geometric coding is completed and the geometric information is reconstructed, color conversion is required to convert the color information (i.e., attribute information) from the RGB color space to the YUV color space. Then, the point cloud is recolored using the reconstructed geometric information so that the uncoded attribute information corresponds to the reconstructed geometric information. Attribute coding is mainly performed on color information.
  • the distance-based lifting transformation that depends on the level of detail (LOD) division
  • RAHT direct region adaptive hierarchical transformation
  • FIG4B shows a schematic diagram of the composition framework of a G-PCC decoder.
  • the geometric bit stream and the attribute bit stream in the binary bit stream are first decoded independently.
  • the geometric information of the point cloud is obtained through arithmetic decoding-reconstruction of the octree/reconstruction of the prediction tree-reconstruction of the geometry-coordinate inverse conversion;
  • the attribute information of the point cloud is obtained through arithmetic decoding-inverse quantization-LOD partitioning/RAHT-color inverse conversion, and the point cloud data to be encoded (i.e., the output point cloud) is restored based on the geometric information and attribute information.
  • the current geometric coding of G-PCC can be divided into octree-based geometric coding (marked by a dotted box) and prediction tree-based geometric coding (marked by a dotted box).
  • the octree-based geometry encoding includes: first, coordinate transformation of the geometric information so that all point clouds are contained in a bounding box. Then quantization is performed. This step of quantization mainly plays a role of scaling. Due to the quantization rounding, the geometric information of some points is the same. The parameters are used to decide whether to remove duplicate points. The process of quantization and removal of duplicate points is also called voxelization. Next, the bounding box is continuously divided into trees (such as octrees, quadtrees, binary trees, etc.) in the order of breadth-first traversal, and the placeholder code of each node is encoded.
  • trees such as octrees, quadtrees, binary trees, etc.
  • a company proposed an implicit geometry partitioning method.
  • the bounding box of the point cloud is calculated. Assume that dx > dy > dz , the bounding box corresponds to a cuboid.
  • K and M In the process of binary tree/quadtree/octree partitioning, two parameters are introduced: K and M.
  • K indicates the maximum number of binary tree/quadtree partitions before octree partitioning;
  • parameter M is used to indicate that the minimum block side length corresponding to binary tree/quadtree partitioning is 2 M.
  • the reason why parameters K and M meet the above conditions is that in the process of geometric implicit partitioning in G-PCC, the priority of partitioning is binary tree, quadtree and octree.
  • the node block size does not meet the conditions of binary tree/quadtree, the node will be partitioned by octree until it is divided into the minimum unit of leaf node 1 ⁇ 1 ⁇ 1.
  • the geometric information encoding mode based on octree can effectively encode the geometric information of point cloud by utilizing the correlation between adjacent points in space.
  • the encoding efficiency of point cloud geometric information can be further improved by using plane coding.
  • Fig. 5A and Fig. 5B provide a kind of plane position schematic diagram.
  • Fig. 5A shows a kind of low plane position schematic diagram in the Z-axis direction
  • Fig. 5B shows a kind of high plane position schematic diagram in the Z-axis direction.
  • (a), (a0), (a1), (a2), (a3) here all belong to the low plane position in the Z-axis direction.
  • the four subnodes occupied in the current node are located at the high plane position of the current node in the Z-axis direction, so it can be considered that the current node belongs to a Z plane and is a high plane in the Z-axis direction.
  • FIG. 6 provides a schematic diagram of the node coding order, that is, the node coding is performed in the order of 0, 1, 2, 3, 4, 5, 6, and 7 as shown in FIG. 6.
  • the octree coding method is used for (a) in FIG. 5A, the placeholder information of the current node is represented as: 11001100.
  • the plane coding method is used, first, an identifier needs to be encoded to indicate that the current node is a plane in the Z-axis direction.
  • the plane position of the current node needs to be represented; secondly, only the placeholder information of the low plane node in the Z-axis direction needs to be encoded (that is, the placeholder information of the four subnodes 0, 2, 4, and 6). Therefore, based on the plane coding method, only 6 bits need to be encoded to encode the current node, which can reduce the representation of 2 bits compared with the octree coding of the related art. Based on this analysis, plane coding has a more obvious coding efficiency than octree coding.
  • FIG7A shows a schematic diagram of plane identification information.
  • PlaneMode _i 0 means that the current node is not a plane in the i-axis direction, and 1 means that the current node is a plane in the i-axis direction. If the current node is a plane in the i-axis direction, then for PlanePosition _i : 0 means that the current node is a plane in the i-axis direction, and the plane position is a low plane, and 1 means that the current node is a high plane in the i-axis direction.
  • the threshold is adaptively changed. For example, when Prob(0)>Prob(1)>Prob(2), the setting of Eligible i is as follows:
  • Prob(i) new (L ⁇ Prob(i)+ ⁇ (coded node))/L+1 (1)
  • L 255; in addition, if the coded node is a plane, ⁇ (coded node) is 1; otherwise, ⁇ (coded node) is 0.
  • FIG8 shows a schematic diagram of the sibling nodes of the current node. As shown in FIG8, the current node is a node filled with slashes, and the nodes filled with grids are sibling nodes, then the number of sibling nodes of the current node is 5 (including the current node itself).
  • planarEligibleK OctreeDepth if (pointCount-numPointCountRecon) is less than nodeCount ⁇ 1.3, then planarEligibleK OctreeDepth is true; if (pointCount-numPointCountRecon) is not less than nodeCount ⁇ 1.3, then planarEligibleKOctreeDepth is false. In this way, when planarEligibleKOctreeDepth is true, all nodes in the current layer are plane-encoded; otherwise, all nodes in the current layer are not plane-encoded, and only octree coding is used.
  • Figure 9 shows a schematic diagram of the intersection of a laser radar and a node.
  • a node filled with a grid is simultaneously passed through by two laser beams (Laser), so the current node is not a plane in the vertical direction of the Z axis;
  • a node filled with a slash is small enough that it cannot be passed through by two lasers at the same time, so the node filled with a slash may be a plane in the vertical direction of the Z axis.
  • the plane identification information and the plane position information may be predictively coded.
  • the predictive encoding of the plane position information may include:
  • the plane position information is divided into three elements: predicted as a low plane, predicted as a high plane, and unpredictable;
  • the spatial distance after determining the spatial distance between the node at the same division depth and the same coordinates as the current node and the current node, if the spatial distance is less than a preset distance threshold, then the spatial distance can be determined to be "near”; or, if the spatial distance is greater than the preset distance threshold, then the spatial distance can be determined to be "far”.
  • FIG10 shows a schematic diagram of neighborhood nodes at the same division depth and the same coordinates.
  • the bold large cube represents the parent node (Parent node), the small cube filled with a grid inside it represents the current node (Current node), and the intersection position (Vertex position) of the current node is shown;
  • the small cube filled with white represents the neighborhood nodes at the same division depth and the same coordinates, and the distance between the current node and the neighborhood node is the spatial distance, which can be judged as "near” or "far”; in addition, if the neighborhood node is a plane, then the plane position (Planar position) of the neighborhood node is also required.
  • the current node is a small cube filled with a grid
  • the neighboring node is searched for a small cube filled with white at the same octree partition depth level and the same vertical coordinate, and the distance between the two nodes is judged as "near" and "far", and the plane position of the reference node is referenced.
  • FIG11 shows a schematic diagram of a current node being located at a low plane position of a parent node.
  • (a), (b), and (c) show three examples of the current node being located at a low plane position of a parent node.
  • the specific description is as follows:
  • FIG12 shows a schematic diagram of a current node being located at a high plane position of a parent node.
  • (a), (b), and (c) show three examples of the current node being located at a high plane position of a parent node.
  • the specific description is as follows:
  • Figure 13 shows a schematic diagram of predictive encoding of the laser radar point cloud plane position information.
  • the laser radar emission angle is ⁇ bottom
  • it can be mapped to the bottom plane (Bottom virtual plane)
  • the laser radar emission angle is ⁇ top
  • it can be mapped to the top plane (Top virtual plane).
  • the plane position of the current node is predicted by using the laser radar acquisition parameters, and the position of the current node intersecting with the laser ray is used to quantify the position into multiple intervals, which is finally used as the context information of the plane position of the current node.
  • the specific calculation process is as follows: Assuming that the coordinates of the laser radar are (x Lidar , y Lidar , z Lidar ), and the geometric coordinates of the current node are (x, y, z), then first calculate the vertical tangent value tan ⁇ of the current node relative to the laser radar, and the calculation formula is as follows:
  • each Laser has a certain offset angle relative to the LiDAR, it is also necessary to calculate the relative tangent value tan ⁇ corr,L of the current node relative to the Laser.
  • the specific calculation is as follows:
  • the relative tangent value tan ⁇ corr,L of the current node is used to predict the plane position of the current node. Specifically, assuming that The tangent value of the lower boundary of the previous node is tan( ⁇ bottom ), and the tangent value of the upper boundary is tan( ⁇ top ). According to tan ⁇ corr,L, the plane position is quantized into 4 quantization intervals, that is, the context information of the plane position is determined.
  • the octree-based geometric information coding mode only has an efficient compression rate for points with correlation in space.
  • the use of the direct coding model (DCM) can greatly reduce the complexity.
  • DCM direct coding model
  • the use of DCM is not represented by flag information, but is inferred from the parent node and neighbor information of the current node. There are three ways to determine whether the current node is eligible for DCM encoding, as follows:
  • the current node has no sibling child nodes, that is, the parent node of the current node has only one child node, and the parent node of the parent node of the current node has only two occupied child nodes, that is, the current node has at most one neighbor node.
  • the parent node of the current node has only one child node, the current node.
  • the six neighbor nodes that share a face with the current node are also empty nodes.
  • FIG14 provides a schematic diagram of IDCM coding. If the current node does not have the DCM coding qualification, it will be divided into octrees. If it has the DCM coding qualification, the number of points contained in the node will be further determined. When the number of points is less than a threshold value (for example, 2), the node will be DCM-encoded, otherwise the octree division will continue.
  • a threshold value for example, 2
  • IDCM_flag the current node is encoded using DCM, otherwise octree coding is still used.
  • the DCM coding mode of the current node needs to be encoded.
  • DCM modes There are currently two DCM modes, namely: (a) only one point exists (or multiple points, but they are repeated points); (b) contains two points.
  • the geometric information of each point needs to be encoded. Assuming that the side length of the node is 2d , d bits are required to encode each component of the geometric coordinates of the node, and the bit information is directly encoded into the bit stream. It should be noted here that when encoding the lidar point cloud, the three-dimensional coordinate information can be predictively encoded by using the lidar acquisition parameters, thereby further improving the encoding efficiency of the geometric information.
  • the current node does not meet the requirements of the DCM node, it will exit directly (that is, the number of points is greater than 2 points and it is not a duplicate point).
  • the second point of the current node is a repeated point, and then it is encoded whether the number of repeated points of the current node is greater than 1. When the number of repeated points is greater than 1, it is necessary to perform exponential Golomb decoding on the remaining number of repeated points.
  • the coordinate information of the points contained in the current node is encoded.
  • the following will introduce the lidar point cloud and the human eye point cloud in detail.
  • the priority coded coordinate axis dirextAxis will be obtained first by using the geometric coordinates of the points. It should be noted here that the coordinate axes currently compared only include the x-axis and the y-axis, but not the z-axis. Assuming that the geometric coordinates of the current node are nodePos, the judgment method is as follows:
  • the axis with the smaller node coordinate geometry position will be used as the priority coded axis dirextAxis, and then the geometry information of the priority coded axis dirextAxis will be encoded as follows. Assume that the bit depth of the coded geometry corresponding to the priority coded axis is nodeSizeLog2, and assume that the coordinates of the two points are pointPos[0] and pointPos[1].
  • the specific encoding process is as follows:
  • the priority coded coordinate axis dirextAxis will be obtained first by using the geometric coordinates of the points. Assuming that the geometric coordinates of the current node are nodePos, the judgment method is as follows:
  • the priority coded coordinate axis dirextAxis geometry information is first encoded as follows, assuming that the priority coded axis corresponds to the coded geometry bit depth of nodeSizeLog2, and assuming that the coordinates of the two points are pointPos[0] and pointPos[1].
  • the specific encoding process is as follows:
  • the geometric coordinate information of the current node can be predicted, so as to further improve the efficiency of the geometric information encoding of the point cloud.
  • the geometric information nodePos of the current node is first used to obtain a directly encoded main axis direction, and then the geometric information of the encoded direction is used to predict the geometric information of another dimension.
  • the axis direction of the direct encoding is directAxis
  • the bit depth of the direct encoding is nodeSizeLog2
  • FIG15 provides a schematic diagram of coordinate transformation of a rotating laser radar to obtain a point cloud.
  • the (x, y, z) coordinates of each node can be converted to (R, i).
  • the laser scanner can perform laser scanning at a preset angle, and different ⁇ (i) can be obtained under different values of i.
  • ⁇ (1) can be obtained, and the corresponding scanning angle is -15°; when i is equal to 2, ⁇ (2) can be obtained, and the corresponding scanning angle is -13°; when i is equal to 10, ⁇ (10) can be obtained, and the corresponding scanning angle is +13°; when i is equal to 9, ⁇ (19) can be obtained, and the corresponding scanning angle is +15°.
  • the LaserIdx corresponding to the current point i.e., the pointLaserIdx number in Figure 15, will be calculated first, and the LaserIdx of the current node, i.e., nodeLaserIdx, will be calculated; secondly, the LaserIdx of the node, i.e., nodeLaserIdx, will be used to predictively encode the LaserIdx of the point, i.e., pointLaserIdx, where the calculation method of the LaserIdx of the node or point is as follows.
  • the LaserIdx of the current node is first used to predict the pointLaserIdx of the point. After the LaserIdx of the current point is encoded, the three-dimensional geometric information of the current point is predicted and encoded using the acquisition parameters of the laser radar.
  • FIG16 shows a schematic diagram of predictive coding in the X-axis or Y-axis direction.
  • a box filled with a grid represents a current node
  • a box filled with a slash represents an already coded node.
  • the LaserIdx corresponding to the current node is first used to obtain the corresponding predicted value of the horizontal azimuth, that is, Secondly, the node geometry information corresponding to the current point is used to obtain the horizontal azimuth angle corresponding to the node Assuming the geometric coordinates of the node are nodePos, the horizontal azimuth
  • the calculation method between the node geometry information is as follows:
  • Figure 17A shows a schematic diagram of predicting the angle of the Y plane through the horizontal azimuth angle
  • Figure 17B shows a schematic diagram of predicting the angle of the X plane through the horizontal azimuth angle.
  • the calculation method is as follows:
  • FIG18 shows another schematic diagram of predictive coding in the X-axis or Y-axis direction.
  • the portion filled with a grid represents the low plane
  • the portion filled with dots represents the high plane.
  • Indicates the horizontal azimuth of the low plane of the current node Indicates the horizontal azimuth of the high plane of the current node, Indicates the predicted horizontal azimuth angle corresponding to the current node.
  • the LaserIdx corresponding to the current point will be used to predict the Z-axis direction of the current point. That is, the depth information radius of the radar coordinate system is calculated by using the x and y information of the current point. Then, the tangent value of the current point and the vertical offset are obtained by using the laser LaserIdx of the current point, and the predicted value of the Z-axis direction of the current point, namely Z_pred, can be obtained.
  • the details are as follows:
  • Z_pred radius ⁇ tanTheta-zOffset.
  • Z_pred is used to perform predictive coding on the geometric information of the current point in the Z-axis direction to obtain the prediction residual Z_res, and finally Z_res is encoded.
  • G-PCC currently introduces a plane coding mode. In the process of geometric division, it will determine whether the child nodes of the current node are in the same plane. If the child nodes of the current node meet the conditions of the same plane, the child nodes of the current node will be represented by the plane.
  • the decoder follows the order of breadth-first traversal. Before decoding the placeholder information of each node, it first uses the reconstructed geometric information to determine whether the current node is to be decoded in plane or IDCM. If If the current node meets the conditions for plane decoding, the plane identification and plane position information of the current node will be decoded first, and then the placeholder information of the current node will be decoded based on the plane information; if the current node meets the conditions for IDCM decoding, it will first decode whether the current node is a true IDCM node.
  • the placeholder information of the current node will be decoded.
  • the placeholder code of each node is obtained, and the nodes are continuously divided in turn until a 1 ⁇ 1 ⁇ 1 unit cube is obtained, the division is stopped, the number of points contained in each leaf node is obtained by parsing, and finally the geometric reconstructed point cloud information is restored.
  • the prior information is first used to determine whether the node starts IDCM. That is, the starting conditions of IDCM are as follows:
  • the current node has no sibling child nodes, that is, the parent node of the current node has only one child node, and the parent node of the parent node of the current node has only two occupied child nodes, that is, the current node has at most one neighbor node.
  • the parent node of the current node has only one child node, the current node.
  • the six neighbor nodes that share a face with the current node are also empty nodes.
  • a node meets the conditions for DCM coding, first decode whether the current node is a real DCM node, that is, IDCM_flag; when IDCM_flag is true, the current node adopts DCM coding, otherwise it still adopts octree coding.
  • numPonts of the current node obtained by decoding is less than or equal to 1, continue decoding to see if the second point is a repeated point; if the second point is not a repeated point, it can be implicitly inferred that the second type that satisfies the DCM mode contains only one point; if the second point obtained by decoding is a repeated point, it can be inferred that the third type that satisfies the DCM mode contains multiple points, but they are all repeated points, then continue decoding to see if the number of repeated points is greater than 1 (entropy decoding), and if it is greater than 1, continue decoding the number of remaining repeated points (decoding using exponential Columbus).
  • the current node does not meet the requirements of the DCM node, it will exit directly (that is, the number of points is greater than 2 points and it is not a duplicate point).
  • the coordinate information of the points contained in the current node is decoded.
  • the following will introduce the lidar point cloud and the human eye point cloud in detail.
  • the geometric coordinates of the points will be used to obtain the priority decoding coordinate axis dirextAxis. It should be noted that the coordinate axes currently compared only include the x and y axes, not the z axis. Assuming that the geometric coordinates of the current node are nodePos, the judgment method is as follows:
  • the axis with the smaller node coordinate geometry position will be used as the priority decoding axis dirextAxis, and then the priority decoding axis dirextAxis geometry information will be decoded first in the following way.
  • the geometry bit depth to be decoded corresponding to the priority decoding axis is nodeSizeLog2
  • the coordinates of the two points are pointPos[0] and pointPos[1] respectively.
  • the specific encoding process is as follows:
  • the geometric coordinates of the points will be used to obtain the priority decoding axis dirextAxis. Assuming that the geometric coordinates of the current node are nodePos, the judgment method is as follows:
  • the priority encoded coordinate axis dirextAxis geometry information is first decoded as follows, assuming that the priority decoded axis corresponds to the code geometry bit depth of nodeSizeLog2, and assuming that the coordinates of the two points are pointPos[0] and pointPos[1].
  • the specific encoding process is as follows:
  • the LaserIdx of the current node i.e., nodeLaserIdx
  • the LaserIdx of the node i.e., nodeLaserIdx
  • the calculation method of the LaserIdx of the node or point is the same as that of the encoder.
  • the LaserIdx of the current point and the predicted residual information of the LaserIdx of the node are decoded to obtain ResLaserIdx.
  • the decoding method is as follows:
  • the three-dimensional geometric information of the current point is predicted and decoded using the acquisition parameters of the laser radar.
  • the specific algorithm is as follows:
  • the node geometry information corresponding to the current point is used to obtain the horizontal azimuth angle corresponding to the node Assuming the geometric coordinates of the node are nodePos, the horizontal azimuth
  • the calculation method between the node geometry information is as follows:
  • the Z-axis direction of the current point will be predicted and decoded using the LaserIdx corresponding to the current point, that is, the depth information radius of the radar coordinate system is calculated by using the x and y information of the current point, and then the tangent value of the current point and the vertical offset are obtained using the laser LaserIdx of the current point, so the predicted value of the Z-axis direction of the current point, namely Z_pred, can be obtained.
  • the details are as follows:
  • Z_pred radius ⁇ tanTheta-zOffset.
  • the decoded Z_res and Z_pred are used to reconstruct and restore the geometric information of the current point in the Z-axis direction.
  • geometric information coding based on triangle soup (trisoup)
  • geometric division must also be performed first, but different from geometric information coding based on binary tree/quadtree/octree, this method does not need to divide the point cloud into unit cubes with a side length of 1 ⁇ 1 ⁇ 1 step by step, but stops dividing when the side length of the sub-block is W.
  • the surface and the twelve edges of the block are obtained.
  • the vertex coordinates of each block are encoded in turn to generate a binary code stream.
  • the Predictive geometry coding includes: first, sorting the input point cloud.
  • the currently used sorting methods include unordered, Morton order, azimuth order, and radial distance order.
  • the prediction tree structure is established by using two different methods, including: KD-Tree (high-latency slow mode) and low-latency fast mode (using laser radar calibration information).
  • KD-Tree high-latency slow mode
  • low-latency fast mode using laser radar calibration information.
  • each node in the prediction tree is traversed, and the geometric position information of the node is predicted by selecting different prediction modes to obtain the prediction residual, and the geometric prediction residual is quantized using the quantization parameter.
  • the prediction residual of the prediction tree node position information, the prediction tree structure, and the quantization parameters are encoded to generate a binary code stream.
  • the decoding end reconstructs the prediction tree structure by continuously parsing the bit stream, and then obtains the geometric position prediction residual information and quantization parameters of each prediction node through parsing, and dequantizes the prediction residual to recover the reconstructed geometric position information of each node, and finally completes the geometric reconstruction of the decoding end.
  • attribute encoding is mainly performed on color information.
  • the color information is converted from RGB color space to YUV color space.
  • the point cloud is recolored using the reconstructed geometric information so that the unencoded attribute information corresponds to the reconstructed geometric information.
  • color information encoding there are two main transformation methods. One is based on The first method is to directly perform RAHT transformation based on distance of LOD division. Both methods will convert color information from spatial domain to frequency domain, obtain high-frequency coefficients and low-frequency coefficients through transformation, and finally quantize and encode the coefficients to generate binary code stream, as shown in Figures 4A and 4B.
  • the Morton code can be used to search for the nearest neighbor.
  • the Morton code corresponding to each point in the point cloud can be obtained from the geometric coordinates of the point.
  • the specific method for calculating the Morton code is described as follows. For each component of the three-dimensional coordinate represented by a d-bit binary number, its three components can be expressed as:
  • the Morton code M is x, y, z starting from the highest bit, and then arranged in sequence from x l ,y l ,z l to the lowest bit.
  • the calculation formula of M is as follows:
  • Condition 1 The geometric position is limitedly lossy and the attributes are lossy;
  • Condition 3 The geometric position is lossless, and the attributes are limitedly lossy
  • Condition 4 The geometric position and attributes are lossless.
  • the general test sequences include four categories: Cat1A, Cat1B, Cat3-fused, and Cat3-frame.
  • the Cat2-frame point cloud only contains reflectance attribute information
  • the Cat1A and Cat1B point clouds only contain color attribute information
  • the Cat3-fused point cloud contains both color and reflectance attribute information.
  • the bounding box is divided into sub-cubes in sequence, and the non-empty sub-cubes (containing points in the point cloud) are divided again until the leaf node obtained by division is a 1 ⁇ 1 ⁇ 1 unit cube.
  • the number of points contained in the leaf node needs to be encoded, and finally the encoding of the geometric octree is completed to generate a binary code stream.
  • the decoding end reconstructs the prediction tree structure by continuously parsing the bit stream, and then obtains the geometric position prediction residual information and quantization parameters of each prediction node through parsing, and dequantizes the prediction residual to restore the reconstructed geometric position information of each node, and finally completes the geometric reconstruction at the decoding end.
  • the current G-PCC coding framework includes three attribute coding methods: Predicting Transform (PT), Lifting Transform (LT), and Region Adaptive Hierarchical Transform (RAHT).
  • PT Predicting Transform
  • LT Lifting Transform
  • RAHT Region Adaptive Hierarchical Transform
  • the first two predict the point cloud based on the generation order of LOD
  • RAHT adaptively transforms the attribute information from bottom to top based on the construction level of the octree.
  • PT Predicting Transform
  • LT Lifting Transform
  • RAHT Region Adaptive Hierarchical Transform
  • the attribute prediction module of G-PCC adopts a nearest neighbor attribute prediction coding scheme based on a hierarchical (Level-of-details, LoDs) structure.
  • the LOD construction methods include distance-based LOD construction schemes, fixed sampling rate-based LOD construction schemes, and octree-based LOD construction schemes.
  • the point cloud is first Morton sorted before constructing the LOD to ensure that there is a strong attribute correlation between adjacent points.
  • Rl point cloud detail layers
  • the attribute value of each point is linearly weighted predicted by using the attribute reconstruction value of the point in the same layer or higher LOD, where the maximum number of reference prediction neighbors is determined by the encoder high-level syntax elements.
  • the encoding end uses the rate-distortion optimization algorithm to select the weighted prediction by using the attributes of the N nearest neighbor points searched or the attribute of a single nearest neighbor point for prediction, and finally encodes the selected prediction mode and prediction residual.
  • N represents the number of predicted points in the nearest neighbor point set of point i
  • Pi represents the sum of the N nearest neighbor points of point i
  • Dm represents the spatial geometric distance from the nearest neighbor point m to the current point i
  • Attrm represents the attribute value after reconstruction of the nearest neighbor point m
  • Attr i ′ represents the attribute prediction value of the current point i
  • the number of points N is a preset value.
  • a switch is introduced in the encoder high-level syntax element to control whether to introduce LOD layer intra prediction. If it is turned on, LOD layer intra prediction is enabled, and points in the same LOD layer can be used for prediction. It should be noted that when the number of LOD layers is 1, LOD layer intra prediction is always used.
  • FIG21 is a schematic diagram of a visualization result of the LOD generation process. As shown in FIG21, a subjective example of the distance-based LOD generation process is provided. Specifically (from left to right): the points in the first layer represent the outer contour of the point cloud; as the number of detail layers increases, the point cloud detail description becomes clearer.
  • Figure 22 is a schematic diagram of the encoding process of attribute prediction.
  • attribute prediction for the specific process of G-PCC attribute prediction, for the original point cloud, first search for the three neighboring points of the Kth point, and then perform attribute prediction; calculate the difference between the attribute prediction value of the Kth point and the original attribute value of the Kth point to obtain the prediction residual of the Kth point; then perform quantization and arithmetic coding to finally generate the attribute bit rate.
  • the LOD After the LOD is constructed, according to the generation order of LOD, first find the three nearest neighbor points of the current point to be encoded from the encoded data points. The attribute reconstruction values of these three nearest neighbor points are used as candidate prediction values of the current point to be encoded; then, the optimal prediction value is selected from them according to the rate-distortion optimization (RDO).
  • RDO rate-distortion optimization
  • the prediction variable index of the attribute value of the nearest neighbor point P4 is set to 1; the attribute prediction variable indexes of the second nearest neighbor point P5 and the third nearest neighbor point P0 are set to 2 and 3 respectively; the prediction variable index of the weighted average of points P0, P5 and P4 is set to 0, as shown in Table 1; finally, use RDO to select the best prediction variable.
  • the formula for weighted average is as follows:
  • x i , y i , zi are the geometric position coordinates of the current point i
  • x ij , y ij , z ij are the geometric coordinates of the neighboring point j.
  • Table 1 provides an example of a sample of candidate prediction items for an attribute encoding.
  • the attribute prediction value of the current point i is obtained through the above prediction (k is the total number of points in the point cloud).
  • (a i ) i ⁇ 0...k-1 be the original attribute value of the current point, then the attribute residual (r i ) i ⁇ 0...k-1 is recorded as:
  • the prediction residuals are further quantified:
  • Qi represents the quantized attribute residual of the current point i
  • Qs is the quantization step (Quantization step, Qs), which can be calculated by the quantization parameter QP (Quantization Parameter, QP) specified by CTC.
  • the purpose of reconstruction at the encoding end is to predict subsequent points. Before reconstructing the attribute value, the residual must be dequantized. is the residual after inverse quantization:
  • intra-frame nearest neighbor search When performing attribute nearest neighbor search based on LOD division, there are currently two major types of algorithms: intra-frame nearest neighbor search and inter-frame nearest neighbor search.
  • inter-frame nearest neighbor search algorithm is as follows, and the intra-frame nearest neighbor search can be divided into two algorithms: inter-layer nearest neighbor search and intra-layer nearest neighbor search.
  • the nearest neighbor search within a frame is divided into two algorithms: the inter-layer nearest neighbor search and the intra-layer nearest neighbor search. After LOD division, it is similar to a pyramid structure, as shown in Figure 23.
  • FIG 24 is a schematic diagram of the LOD construction process for the inter-layer nearest neighbor search.
  • Figure 25 is a schematic diagram of the LOD construction process for the inter-layer nearest neighbor search. As shown in Figure 25, different LOD layers are obtained based on the geometric information division, and LOD0, LOD1 and LOD2 are obtained. The points in LOD0 are used to predict the attributes of the points in the next layer of LOD in the process of the inter-layer nearest neighbor search.
  • the entire LOD division process there are three sets O(k), L(k) and I(k). Among them, k is the index of the LOD layer during LOD division, I(k) is the input point set during the current LOD layer division, and after LOD division, O(k) set and L(k) set are obtained. The O(k) set stores the sampling point set, and L(k) is the point set in the current LOD layer. That is, the entire LOD division process is as follows:
  • O(k), L(k) and I(k) store the Morton code index corresponding to the point.
  • the neighbor search is performed by using the parent block (Block B) corresponding to point P, as shown in Figure 26, and the points in the neighbor blocks that are coplanar and colinear with the current parent block are searched for attribute prediction.
  • FIG. 27A shows a schematic diagram of a coplanar spatial relationship, where there are 6 spatial blocks that have a relationship with the current parent block.
  • FIG. 27B shows a schematic diagram of a coplanar and colinear spatial relationship, where there are 18 spatial blocks that have a relationship with the current parent block.
  • FIG. 27C shows a schematic diagram of a coplanar, colinear and co-point spatial relationship, where there are 26 spatial blocks that have a relationship with the current parent block.
  • the coordinates of the current point are used to obtain the corresponding spatial block.
  • the nearest neighbor search is performed in the previously encoded LOD layer to find the spatial blocks that are coplanar, colinear, and co-point with the current block to obtain the N nearest neighbors of the current point.
  • the N nearest neighbors of the current point After searching for coplanar, colinear, and co-point nearest neighbors, if the N nearest neighbors of the current point are still not found, the N nearest neighbors of the current point will be found based on the fast search algorithm.
  • the specific algorithm is as follows:
  • the geometric coordinates of the current point to be encoded are first used to obtain the Morton code corresponding to the current point. Secondly, based on the Morton code of the current point, the first reference point (j) that is larger than the Morton code of the current point is found in the reference frame. Then, the nearest neighbor search is performed in the range of [j-searchRange, j+searchRange].
  • a video frame can be understood as an image.
  • the current frame can be understood as the current image
  • the reference frame can be understood as the reference image.
  • FIG29 shows a schematic diagram of the LOD structure of the nearest neighbor search within an attribute layer.
  • the nearest neighbor point of the current point P6 can be P1.
  • the nearest neighbor point can be P4. That is to say, when the intra-layer prediction algorithm is turned on, the nearest neighbor search will be performed in the same layer LOD and the set of encoded points in the same layer to obtain the N nearest neighbors of the current point (the inter-layer nearest neighbor search is also performed).
  • the nearest neighbor search is performed based on the fast search algorithm.
  • the specific algorithm is shown in Figure 30.
  • the current point is represented by a grid.
  • the nearest neighbor search is performed in [i+1, i+searchRange].
  • the specific nearest neighbor search algorithm is consistent with the inter-frame block-based fast search algorithm and will not be described in detail here.
  • Figure 28 is a schematic diagram of attribute inter-frame prediction.
  • attribute inter-frame prediction when performing attribute inter-frame prediction, firstly, the geometric coordinates of the current point to be encoded are used to obtain the Morton code corresponding to the current point, and then the first reference point (j) with a value greater than the Morton code of the current point is found in the reference frame based on the Morton code of the current point, and then the nearest neighbor search is performed within the range of [j-searchRange, j+searchRange].
  • the specific division algorithm is as follows:
  • the reference range in the prediction frame of the current point is [j-searchRange, j+searchRange], use j-searchRange to calculate the starting index of the third layer, and use j+searchRange to calculate the ending index of the third layer; secondly, first determine whether some blocks in the second layer need to be searched for the nearest neighbor in the blocks of the third layer, and then go to the second layer, and determine whether a search is needed for each block in the first layer. If some blocks in the first layer need to be searched for the nearest neighbor, then some midpoints of some blocks in the first layer will be judged point by point to update the nearest neighbors.
  • the index of the first layer block is obtained based on the index of the second layer block based on the same algorithm.
  • MinPos represents the minimum value of the block
  • maxPos represents the maximum value of the block.
  • the coordinates of the point to be encoded are (x, y, z), and the current block is represented by (minPos, maxPos), where minPos is the minimum value of the bounding box in three dimensions, and maxPos is the maximum value of the bounding box in three dimensions.
  • the distance D between the current point and the bounding box is calculated as follows:
  • int dx int(std::max(std::max(minPos[0]-point[0],0),point[0]-maxPos[0]));
  • int dy int(std::max(std::max(minPos[1]-point[1],0),point[1]-maxPos[1]));
  • Figure 32 is a schematic diagram of the encoding process of a lifting transformation.
  • the lifting transformation also predicts the attributes of the point cloud based on LOD.
  • the difference from the prediction transformation is that the lifting transformation first divides the LOD into high and low layers, predicts in the reverse order of the LOD generation layer, and introduces an update operator in the prediction process to update the quantization weights of the midpoints of the low-level LOD to improve the prediction accuracy. Accuracy. This is because the attribute values of points in lower LODs are frequently used to predict the attribute values of points in higher LODs, and points in lower LODs should have greater influence.
  • Step 1 Segmentation process.
  • Step 2 Prediction process.
  • the points in the high-level LOD select the attribute information of the nearest neighbor points from the low-level as the attribute prediction value P(N) of the current point to be encoded, and the prediction residual D(N) is recorded as:
  • Step 3 Update Process.
  • the transformation scheme based on lifting wavelet transform introduces quantization weights and updates the prediction residual according to the prediction residual D(N) and the distance between the prediction point and the adjacent points, and finally uses the quantization weights in the transformation process to adaptively quantize the prediction residual.
  • the quantization weight value of each point can be determined by geometric reconstruction at the decoding end, so the quantization weight should not be encoded.
  • Regional Adaptive Hierarchical Transform is a Haar wavelet transform that can transform point cloud attribute information from the spatial domain to the frequency domain, further reducing the correlation between point cloud attributes. Its main idea is to transform the nodes in each layer from the three dimensions of X, Y, and Z in a bottom-up manner according to the octree structure (as shown in Figure 34), and iterate until the root node of the octree. As shown in Figure 33, its basic idea is to perform wavelet transform based on the hierarchical structure of the octree, associate attribute information with the octree nodes, and recursively transform the attributes of the occupied nodes in the same parent node in a bottom-up manner.
  • RAHT Regional Adaptive Hierarchical Transform
  • the nodes are transformed from the three dimensions of X, Y, and Z until they are transformed to the root node of the octree.
  • the low-pass/low-frequency (DC) coefficients obtained after the transformation of the nodes in the same layer are passed to the nodes in the next layer for further transformation, and all high-pass/high-frequency (AC) coefficients can be encoded by the arithmetic encoder.
  • FIG35A is a schematic diagram of a RAHT forward transformation process
  • FIG35B is a schematic diagram of a RAHT inverse transformation process.
  • g′ L,2x,y,z and g′ L,2x+1,y,z are two attribute DC coefficients of neighboring points in the L layer.
  • the information of the L-1 layer is the AC coefficient f′ L-1,x,y,z and the DC coefficient g′ L-1,x,y,z ; then, f′ L-1,x,y,z will no longer be transformed and will be directly quantized and encoded, and g′ L-1,x,y,z will continue to look for neighbors for transformation.
  • the transformation matrix will be updated as the weights corresponding to each point change adaptively.
  • the above process will be iteratively updated according to the partition structure of the octree until the root node of the octree.
  • prediction can be performed based on RAHT transform coding.
  • the RAHT attribute transform is based on the order of the octree hierarchy, and the transformation is continuously performed from the voxel level until the root node is obtained, thereby completing the hierarchical transform coding of the entire attribute.
  • the attribute prediction transform coding is also performed based on the hierarchical order of the octree, but the transformation is continuously performed from the root node to the voxel level.
  • the attribute prediction transform coding is performed based on a 2 ⁇ 2 ⁇ 2 block. The specific example is shown in Figure 36.
  • the grid filling block is the current block to be encoded
  • the diagonal filling block is some neighboring blocks that are coplanar and colinear with the current block to be encoded.
  • the attributes of the current block are normalized in the following way:
  • a node ⁇ p ⁇ node attribute(p);
  • a node A node /w node .
  • the attributes of the current block can be obtained by the attributes of the points in the current block, that is, A node .
  • the attributes of the points in the current block are simply added, and then the attributes of the current block and the number of points in the current block are normalized to obtain the mean value of the attributes of the current block a node .
  • the mean value of the attributes of the current block is used for attribute transformation coding. For the specific coding process, see Figure 37.
  • RAHT attribute prediction transform coding As shown in Figure 37, the overall process of RAHT attribute prediction transform coding is shown here. Among them, (a) is the current block and some coplanar and colinear neighboring blocks, (b) is the block after normalization, (c) is the block after upsampling, (d) is the attribute of the current block, and (e) is the attribute of the predicted block obtained by linear weighted fitting using the neighborhood attributes of the current block. Finally, the attributes of the two will be transformed respectively to obtain DC and AC coefficients, and the AC coefficient will be predicted and coded.
  • the predicted attribute of the current block can be obtained by linear fitting as shown in FIG38.
  • FIG38 firstly, 19 neighboring blocks of the current block are obtained, and then the attribute of each sub-block is linearly weighted predicted using the spatial geometric distance between the neighboring block and each sub-block of the current block, and finally the predicted block attribute obtained by linear weighting is transformed.
  • the specific attribute transformation is shown in FIG39.
  • (d) represents the original value of the attribute
  • the corresponding attribute transformation coefficient is as follows:
  • (e) represents the attribute prediction value, and the corresponding attribute transformation coefficient is as follows:
  • the prediction residual By subtracting the original value of the attribute from the predicted value of the attribute, the prediction residual can be obtained as follows:
  • the process is similar to the intra-frame prediction coding.
  • the RAHT attribute transformation coding structure is constructed based on the geometric information, that is, the voxel level is continuously transformed until the root node is obtained, thereby completing the hierarchical transformation coding of the entire attribute.
  • the intra-frame coding structure and the inter-frame coding structure are constructed.
  • the inter-frame coding structure of the RAHT attribute can be seen in Figure 40.
  • the geometric information of the current node to be encoded is used to obtain the co-located prediction node of the node to be encoded in the reference frame, and then the geometric information and attribute information of the reference node are used to obtain the predicted attribute of the current node to be encoded.
  • the attribute prediction value of the current node to be encoded is obtained according to the following two different methods:
  • the inter-frame prediction node of the current node is valid: that is, if the same-position node exists, the attribute of the prediction node is directly used as the attribute prediction value of the current node to be encoded;
  • the inter-frame prediction node of the current node is invalid: that is, the co-located node does not exist, then the attribute prediction value of the adjacent node in the frame is used as the attribute prediction value of the node to be encoded.
  • the obtained attribute prediction value is used to predict the attribute of the current node to be encoded, thereby completing the prediction coding of the entire attribute.
  • the RAHT attribute transformation coding structure will be first constructed based on the geometric information of the current node to be encoded, that is, the nodes will be continuously merged at the voxel level until the root node of the entire RAHT transformation tree is obtained, thereby completing the transformation coding hierarchical structure of the entire attribute.
  • the root node is divided to obtain N child nodes (N is less than or equal to 8) of each node.
  • the attributes of the N child nodes will first be independently orthogonally transformed using the RAHT transformation to obtain the DC coefficient and the AC coefficient. Then, the AC coefficients of the N child nodes are predicted for the attribute inter-frame in the following manner:
  • the inter-frame prediction node of the current node is valid: that is, if the same-position node exists, the attribute of the prediction node is directly used as the attribute prediction value of the current node to be encoded
  • the current node can find a node with exactly the same position as the current node in the cache of the reference frame: that is, if the same-position node exists, the AC coefficients of the M child nodes contained in the same-position node are directly used as the AC coefficient attribute prediction values of the N child nodes of the current node.
  • the AC coefficient of the prediction node is not zero: the AC coefficient of the prediction node is directly used as the prediction value;
  • the AC coefficient of the prediction node is zero, the AC coefficient of the corresponding child node of the intra-frame prediction will be used as the prediction value.
  • the inter-frame prediction node of the current node is invalid: that is, the co-located node does not exist, then the attribute prediction value of the adjacent node in the frame is used as the attribute prediction value of the node to be encoded.
  • encoding and decoding can be performed in the order from the root node to the child node.
  • the geometric information of the current layer node is used to restore the child nodes of the current layer in the order of Z, Y and X.
  • the attributes of the current layer node that have been reconstructed are used to predict and decode, thereby restoring the attributes of the current layer node until the transformation is to the child node, that is, the voxel level.
  • the number of nodes in the current layer is exactly the same as the number of child nodes of the current layer node, it means that each node in the current layer has only one child node, that is, the current layer will not generate AC coefficients.
  • the nodes of the current layer still need to be transformed, predicted, and other processes in sequence, which will increase the complexity of RAHT attribute transform encoding and decoding and introduce redundant operations.
  • the codec first completes the encoding and decoding of non-node layer attribute information (i.e., the size of the node is greater than or equal to 1 ⁇ 1 ⁇ 1), and finally encodes and decodes the attribute information of the voxel-level nodes.
  • non-node layer attribute information i.e., the size of the node is greater than or equal to 1 ⁇ 1 ⁇ 1
  • the codec first completes the encoding and decoding of non-node layer attribute information (i.e., the size of the node is greater than or equal to 1 ⁇ 1 ⁇ 1), and finally encodes and decodes the attribute information of the voxel-level nodes.
  • an embodiment of the present application provides an encoding method, first determining a first number of voxel nodes of a current unit and a second number of reconstruction nodes of the current unit; then determining whether to skip encoding the attribute information of the voxel nodes of the current unit based on the first number and the second number.
  • An embodiment of the present application also provides a decoding method, first determining a first number of voxel nodes of a current unit and a second number of reconstruction nodes of the current unit; wherein the first number and the second number are used to determine whether to skip decoding the voxel nodes of the current unit; then determining the attribute reconstruction value of the voxel nodes of the current unit based on the first number and the second number.
  • the judgment conditions for whether to perform attribute encoding and decoding for each voxel node are optimized. Specifically, if there are no duplicate nodes in the current unit, that is, the first number is the same as the second number, there is no need to encode and decode the voxel nodes of the current unit. Therefore, while ensuring the encoding and decoding efficiency of the point cloud attributes, the time complexity of encoding and decoding of the point cloud attributes can be reduced, and the bit rate can also be saved, thereby improving the encoding and decoding performance of the point cloud.
  • FIG41 a schematic diagram of a decoding method provided by an embodiment of the present application is shown. As shown in FIG41, the method may include:
  • S4101 Determine a first number of voxel nodes of a current unit and a second number of reconstruction nodes of the current unit; wherein the first number and the second number are used to determine whether to skip decoding of the voxel nodes of the current unit.
  • the decoding method of the embodiment of the present application is applied to a point cloud decoder (which may be referred to as a "decoder" for short).
  • the method may refer to a point cloud decoding method, specifically a point cloud attribute decoding method, and more specifically, a skip decoding method for point cloud attribute RAHT transform prediction.
  • the optimization is mainly to determine whether the nodes of the current layer are skipped for decoding and whether the voxel nodes are skipped for decoding when divided into voxel levels, thereby reducing the time complexity of attribute transformation encoding and decoding, and will not have any impact on the encoding and decoding efficiency of the attribute transformation.
  • the current unit may be a current decoding unit to be decoded, which may be a slice.
  • the method may further include: dividing the nodes in the current unit to determine at least one layer; when the nodes of the last layer are divided into voxel levels, determining the voxel nodes of the current unit and the first number of voxel nodes.
  • the order of RAHT attribute transformation is to divide from the root node in sequence until it is divided into the voxel level, specifically, the division is stopped when it is divided into a unit cube of size 1 ⁇ 1 ⁇ 1, thereby completing the encoding and reconstruction of the entire point cloud attribute.
  • the layer obtained by downsampling along the Z direction, Y direction and X direction each time is a RAHT transformation layer, that is, layer. Then until it is divided into a unit cube of size 1 ⁇ 1 ⁇ 1, it means that it has been divided into the voxel level, at this time the voxel node and the first number of voxel nodes can be determined.
  • the voxel node represents the node corresponding to the division from the root node to the voxel level, and the size of the voxel node is 1 ⁇ 1 ⁇ 1;
  • the reconstruction node represents the node in the current unit that performs attribute reconstruction.
  • the geometric information of the nodes in the current unit has been fully decoded.
  • the second number of reconstructed nodes in the current unit can be determined, so that it can be determined whether the first number is consistent with the second number.
  • the voxel node before the voxel node is attributed and decoded, it is first necessary to determine whether the first number is the same as the second number, and then determine whether the voxel node of the current unit is skipped for decoding.
  • the first number is the same as the second number, decoding of the voxel node of the current unit is skipped.
  • attribute decoding is performed on the repeated nodes in the voxel nodes, and decoding of the remaining voxel nodes except the repeated nodes in the voxel nodes is skipped.
  • duplicate nodes there may be duplicate nodes in the voxel nodes of the current unit division, so the first number and the second number may be inconsistent. If there are no duplicate nodes in the voxel nodes of the current unit division, the first number and the second number are consistent; conversely, if there are duplicate nodes in the voxel nodes of the current unit division, the first number and the second number are inconsistent. In this case, it is necessary to Attribute decoding for duplicate nodes.
  • duplicate nodes also referred to as "duplicate points” refer to multiple nodes with the same geometric information but different attribute information.
  • S4102 Determine the attribute reconstruction value of the voxel node of the current unit according to the first quantity and the second quantity.
  • determining the attribute reconstruction value of the voxel node of the current unit based on the first number and the second number may include: setting the attribute reconstruction value of the voxel node of the current unit to the attribute reconstruction value of the reconstruction node of the current unit.
  • determining the attribute reconstruction value of the voxel node of the current unit based on the first number and the second number may include: decoding the code stream to determine the attribute reconstruction value of the repeated node; using the attribute reconstruction value of the repeated node as the attribute reconstruction value of the first reconstruction node in the current unit, and setting the attribute reconstruction value of the remaining voxel nodes in the voxel node except the repeated node as the attribute reconstruction value of the remaining reconstruction nodes in the current unit except the first reconstruction node.
  • decoding a code stream and determining attribute reconstruction values of repeated nodes may include: setting a timer; starting the timer when the attribute reconstruction values of repeated nodes begin to be decoded, and determining that all attribute reconstruction values of repeated nodes have been decoded when the timer reaches a preset value.
  • the third number of repeated nodes in the voxel node can also be determined based on the difference between the first number and the second number; wherein the setting of the timer is associated with the third number.
  • a timer may be used to determine whether all attribute reconstruction values of repeated nodes have been decoded. If the timer is in a positive counting mode, then when the attribute reconstruction value of the repeated node begins to be decoded, the initial value of the timer is 0, and when the timer counts to a third number, all attribute reconstruction values of the repeated node are decoded; if the timer is in a countdown mode, then when the attribute reconstruction value of the repeated node begins to be decoded, the initial value of the timer is the third number, and when the timer counts to 0, all attribute reconstruction values of the repeated node are decoded.
  • FIG43 is a flow chart of another decoding method provided in the embodiment of the present application. As shown in FIG43, the method may include:
  • S4301 Determine a fourth number of nodes in a current layer and a fifth number of child nodes corresponding to the nodes in the current layer; wherein the fourth number and the fifth number are used to determine whether to skip decoding of the current layer.
  • a RAHT attribute transformation structure based on the geometric information of the points in the point cloud, and the decoding can be performed in the order from the root node to the child node.
  • the geometric information of the current layer node is used to restore the child nodes of the current layer in the order of Z, Y and X, and then the attributes of the nodes of the previous layer are used to predict and decode the attributes of the nodes of the current layer, so as to restore the attribute reconstruction value of the nodes of the current layer until the transformation is to the voxel level, and then a RAHT attribute transformation structure including at least one RAHT transformation layer can be obtained.
  • the RAHT attribute transformation can be performed based on the order of the octree hierarchy. Based on the hierarchical order of the octree, the transformation can be continuously performed from the voxel level to the root node, thereby constructing the octree. During the transformation process, attribute prediction transformation coding is also performed based on the hierarchical order of the octree, but the root node is continuously transformed until the voxel level.
  • a layer obtained by downsampling in a preset direction is a RAHT transformation layer, such as the current layer.
  • the current layer may include at least one point.
  • the at least one point in the current layer when decoding the current layer, it can be used as a node to be decoded in the current layer.
  • each point in the current layer corresponds to a geometric information and an attribute information; wherein the geometric information represents the spatial relationship of the point, and the attribute information represents the relevant information of the attribute of the point.
  • the attribute information may be color information, or reflectivity or other attributes, which is not specifically limited in the embodiments of the present application.
  • the attribute information may be color information in any color space.
  • the attribute information may be color information in an RGB space, or may be color information in a YUV space, or may be color information in a YCbCr space, etc., which is not specifically limited in the embodiments of the present application.
  • the non-voxel level node attribute transformation and inverse transformation are completed first, and then the voxel level node transformation is completed, because there may be a situation in the point cloud, that is, there are duplicate nodes in the point cloud.
  • the fourth number can represent the number of occupied nodes of the current layer; and the fifth number can represent the number of occupied sub-nodes in the nodes of the current layer.
  • the fourth number can represent the number of occupied nodes of the current layer; and the fifth number can represent the number of nodes to be decoded.
  • the geometric information of the nodes of the current layer may be determined first; and then the child nodes corresponding to the nodes of the current layer and the fifth quantity may be determined based on the geometric information.
  • the current node of the current layer when using the geometric information of the current node to determine the corresponding child node, you can choose to use the geometric information of the current node for upsampling to obtain the child nodes occupied by the current node (the number of child nodes is N, where the maximum value of N is 8).
  • the number of nodes of the current layer when encoding and decoding the attribute information of the nodes of the current layer, the number of nodes of the current layer, that is, the fourth number, can be obtained first; at the same time, after the child nodes of the nodes of the current layer are restored using the geometric information of the nodes of the current layer, the number of child nodes of the nodes of the current layer, that is, the fifth number, can be obtained.
  • S4302 Determine the attribute reconstruction value of the child node corresponding to the node of the current layer according to the fourth quantity and the fifth quantity.
  • the attribute reconstruction value of the child node corresponding to the node of the current layer can be further determined based on the fourth quantity and the fifth quantity.
  • the fourth number and the fifth number can be used to encode and decode the attribute information of whether to skip the coding layer (current layer).
  • the present application by using the number of nodes and the number of child nodes of the current layer, it is possible to adaptively determine whether the current layer can skip encoding and decoding.
  • the key to determining whether to skip encoding and decoding processing of the current layer is whether the number of nodes and the number of child nodes of the current layer are the same, that is, whether the fourth number and the fifth number are the same.
  • the method may further include: if the fourth quantity and the fifth quantity are the same, determining the attribute reconstruction value of the node of the current layer as the attribute reconstruction value of the child node corresponding to the node of the current layer.
  • the RAHT transform is only valid for nodes with neighboring points, during the RAHT attribute transformation process, if each node in the current layer corresponds to only one child node, then it can be considered that the current layer will not generate AC coefficients. Therefore, it can be chosen not to transform, predict, etc. the nodes of the current layer in sequence, that is, skip processing the current layer. At this time, it can be called "skipping the decoding layer.”
  • the fourth quantity and the fifth quantity corresponding to the nodes of the current layer are the same, that is, it is determined that the current layer is a skip decoding layer, then it is possible to choose to skip the transformation, prediction, and other processes of the nodes of the current layer in sequence, and instead directly determine the attribute reconstruction value of the node of the current layer as the attribute reconstruction value of the child node corresponding to the node of the current layer.
  • the fourth number and the fifth number are the same, and the current layer is determined to be a skip decoding layer, then you can choose to skip the current layer to the next layer, and then use the next layer as the current layer, and continue to determine whether to skip decoding the nodes of the next layer.
  • the sixth number of the next layer of child nodes corresponding to the child node can be determined first; then, based on the fifth number and the sixth number, the attribute reconstruction value of the next layer of child nodes corresponding to the child node can be determined.
  • the fifth number and the sixth number are the same, then it may be possible to choose not to perform transformation, prediction, etc. on the child nodes of the node of the current layer in sequence, that is, skip processing the child nodes of the current layer, and instead directly determine the attribute reconstruction value of the child node corresponding to the node of the current layer as the attribute reconstruction value of the next layer of child nodes of the child node corresponding to the node of the current layer.
  • the method may also include: if the fourth quantity and the fifth quantity corresponding to the node of the current layer are different, determining the attribute prediction value of the child node corresponding to the node of the current layer according to the node of the current layer; performing a RAHT transform based on the attribute prediction value of the child node to determine the reconstruction value of the high-frequency coefficient and the low-frequency coefficient corresponding to the node of the current layer; performing a RAHT inverse transform based on the reconstruction value of the high-frequency coefficient and the low-frequency coefficient to determine the attribute reconstruction value of the child node.
  • the current layer can be used as a non-skipped decoding layer.
  • when determining the attribute prediction value of the child node corresponding to the node of the current layer based on the node of the current layer it can include: determining the adjacent nodes corresponding to the node of the current layer; and determining the attribute prediction value of the child node corresponding to the node of the current layer based on the attribute reconstruction value and relative distance parameter corresponding to the adjacent nodes.
  • the current node includes two sub-nodes, sub-node 1 and sub-node 2, and the relative distance parameter between the current node and the adjacent node may include the spatial geometric distance between sub-node 1 and the adjacent node, and may also include the spatial geometric distance between sub-node 2 and the adjacent node.
  • the reconstructed attributes (attribute reconstruction values) of the neighboring nodes of the current node and the spatial geometric distance of each neighboring node from the child node of the current node can be used to perform linear fitting, and finally obtain the attribute prediction value of each child node of the current node.
  • the 19 adjacent nodes of the current node can be determined first, and then the attributes of each child node can be linearly weighted predicted using the spatial geometric distance between the adjacent nodes and each child node of the current node, ultimately obtaining the attribute prediction value of each child node.
  • when performing a RAHT transform based on the attribute prediction value of the child node to determine the reconstructed value of the high-frequency coefficient and the low-frequency coefficient corresponding to the node of the current layer it can include: performing a RAHT transform based on the attribute prediction value of the child node to determine the predicted value of the high-frequency coefficient and the low-frequency coefficient corresponding to the node of the current layer; and determining the reconstructed value of the high-frequency coefficient corresponding to the node of the current layer according to the predicted value of the high-frequency coefficient.
  • the attribute prediction value of the corresponding child node can be used to perform RAHT attribute transformation, so as to obtain the corresponding DC coefficient and AC coefficient, that is, to obtain the DC coefficient and AC coefficient corresponding to the current node.
  • the DC coefficient is the low-frequency coefficient
  • the AC coefficient is the high-frequency coefficient.
  • the AC coefficient obtained by performing RAHT attribute transformation using the attribute prediction value corresponding to the sub-node can be understood as the prediction value of the high-frequency coefficient corresponding to the current node.
  • the determining the reconstructed value of the high-frequency coefficient corresponding to the node of the current layer according to the predicted value of the high-frequency coefficient it can include: decoding the code stream to determine the quantized coefficient residual corresponding to the node of the current layer; and determining the reconstructed value of the high-frequency coefficient corresponding to the node of the current layer according to the predicted value of the high-frequency coefficient and the quantized coefficient residual.
  • the determining the reconstructed value of the high-frequency coefficient corresponding to the node of the current layer based on the predicted value of the high-frequency coefficient and the quantized coefficient residual it can include: dequantizing the quantized coefficient residual to determine the dequantized residual value corresponding to the node of the current layer; determining the reconstructed value of the high-frequency coefficient corresponding to the node of the current layer based on the dequantized residual value corresponding to the node of the current layer and the predicted value of the high-frequency coefficient corresponding to the node of the current layer.
  • the inverse quantized residual value corresponding to the node of the current layer and the predicted value of the high-frequency coefficient corresponding to the node of the current layer can be summed up to obtain the reconstructed value of the high-frequency coefficient corresponding to the node of the current layer.
  • a RAHT inverse transform can be performed based on the reconstruction values of the high-frequency coefficients and the low-frequency coefficients, and then the attribute reconstruction values of the child nodes can be determined.
  • g′ L,2x,y,z and g′ L,2x+1,y,z are two attribute DC coefficients of neighboring points in the L layer.
  • the information of the L-1 layer is the AC coefficient f′ L-1,x,y,z and the DC coefficient g′ L-1,x,y,z ; then, f′ L-1,x,y,z will no longer be transformed and will be directly quantized and encoded, and g′ L-1,x,y,z will continue to look for neighbors for transformation.
  • the weights (the number of non-empty child nodes in the node) corresponding to g′ L,2x,y,z and g′ L,2x+2,y ,z are w′ L,2x,y,z and w′ L,2x+1,y,z (abbreviated as w′ 0 and w′ 1 ) respectively, and the weight of g′ L-1,x,y,z is w′ L-1,x,y,z .
  • the general transformation formula is:
  • T w0,w1 is a transformation matrix, and the transformation matrix will be updated as the weights corresponding to each point change adaptively.
  • the forward transformation of RAHT (also referred to as "RAHT forward transformation") is shown in the aforementioned FIG. 35A.
  • the inverse transformation of RAHT is performed according to the DC coefficient and AC coefficient of the child node of the current node, and the attribute reconstruction value of the child node of the current node can be restored.
  • the inverse transformation of RAHT (also referred to as "RAHT inverse transformation” or "RAHT inverse transformation”) is shown in the aforementioned FIG. 35B.
  • the nodes of the current layer can continue to be transformed, predicted, and the like in sequence.
  • the reconstruction attributes of the neighboring nodes of the current node and the spatial geometric distance of each neighboring node from each child node of the current node can be used for linear fitting to obtain the predicted attributes of each child node of the current node; then, the predicted attributes of each child node are used to perform RAHT attribute transformation to obtain the corresponding DC and AC coefficients, and then the AC coefficient of the predicted node (the predicted value of the high-frequency coefficient) and the AC coefficient (coefficient difference) parsed from the bitstream are used to restore the AC coefficient (the reconstructed value of the high-frequency coefficient) of the current node to be decoded (the current node), and finally, the AC coefficient (the reconstructed value of the high-frequency coefficient) and the DC coefficient of
  • the nodes of the current layer can be transformed, predicted, and the like in sequence to determine the attribute reconstruction values of the child nodes corresponding to the nodes of the current layer; then, for the child nodes of the current layer nodes, the sixth number of the next-layer child nodes corresponding to the child nodes can be determined first; and then, based on the fifth number and the sixth number, the attribute reconstruction values of the next-layer child nodes corresponding to the child nodes can be determined.
  • step S4301 to step S4302 is continuously repeated, starting from the root node of the RAHT transform until the last node of the leaf node layer of the RAHT, thereby completing the attribute decoding of the entire RAHT transform.
  • the method may also include: decoding the code stream, determining prediction mode identification information; when the prediction mode identification information indicates that the current unit starts the skip decoding mode, executing the first quantity and the second quantity determination steps, and/or executing the fourth quantity and the fifth quantity determination steps.
  • the prediction mode identification information is at least one of the following high-level syntax elements: a syntax element corresponding to an attribute parameter set (Attribute Parameter Set, APS) and a syntax element corresponding to an attribute block header information (Attribute Block Head, ABH).
  • the value of the prediction mode identification information is the first value, it is determined that the current unit starts the skip decoding mode; if the value of the prediction mode identification information is the second value, it is determined that the current unit does not start the skip decoding mode.
  • the following describes whether the voxel node of the current unit starts the skip decoding mode and whether the node of the current layer starts the skip decoding mode.
  • the method may further include: decoding the code stream to determine first prediction mode identification information; when the first prediction mode identification information indicates that the voxel node of the current unit starts the skip decoding mode, executing the first quantity and the second quantity determination steps.
  • the value of the first prediction mode identification information is a first value, it is determined that the voxel node of the current unit starts the skip decoding mode; if the value of the first prediction mode identification information is a second value, it is determined that the voxel node of the current unit does not start the skip decoding mode.
  • the first number and the second number can be further determined at this time, and then the voxel node of the current unit is determined to skip decoding according to the size of the first number and the second number. Specifically, if the two are consistent, the node attributes at the voxel level are skipped for decoding, otherwise, a timer is used.
  • the number of remaining repeated nodes of the timer is zero, it means that there are no repeated nodes in the subsequent points, and the attribute decoding of the subsequent points can also be skipped, and the attribute reconstruction value of the subsequent voxel node is directly copied as the attribute reconstruction value of the final remaining reconstruction point.
  • the method may also include: decoding the code stream to determine the second prediction mode identification information; when the second prediction mode identification information indicates that the node of the current layer starts the skip decoding mode, executing the fourth quantity and the fifth quantity determination steps.
  • the value of the second prediction mode identification information is a first value, it is determined that the node of the current layer starts the skip decoding mode; if the value of the second prediction mode identification information is a second value, it is determined that the node of the current layer does not start the skip decoding mode.
  • the fourth number and the fifth number can be further determined at this time, and then the size of the fourth number and the fifth number is used to determine whether the node of the current layer skips decoding, that is, whether the current layer is a skip decoding layer. Specifically, if the two are consistent, the current layer is determined to be a skip decoding layer, and it is no longer necessary to decode the attribute reconstruction value of the child node corresponding to the node of the current layer; if the two are inconsistent, it is determined that the current layer does not belong to the skip decoding layer, and RAHT prediction and decoding can be performed according to the decoding method of the relevant technology.
  • the first value is different from the second value, and the first value and the second value can be in parameter form or in digital form.
  • the first prediction mode identification information and the second prediction mode identification information can be parameters written in the profile, or can be the value of a flag, which is not specifically limited here.
  • the first value can be set to 1 and the second value can be set to 0; or, the first value can be set to 0 and the second value can be set to 1; or, the first value can be set to true and the second value can be set to false; or, the first value can be set to false and the second value can be set to true.
  • the first value is set to 1 and the second value is set to 0, but it is not specifically limited.
  • the code stream is decoded to determine the value of the second prediction mode identification information; if the value of the second prediction mode identification information is 1, then it can be determined that the node of the current layer starts the skip decoding mode, and then the fourth quantity and the fifth quantity corresponding to the node of the current layer can be further determined according to the above method; if the value of the second prediction mode identification information is 0, then it can be determined that the node of the current layer does not start the skip decoding mode, and the attribute decoding processing of the node of the current layer can be performed according to the common intra-frame prediction method or inter-frame prediction method.
  • the current layer when encoding and decoding attribute information, if the number of nodes in the current layer is consistent with the number of child nodes in the current layer, the current layer is considered to belong to a skip coding layer, and therefore there is no need to perform transformation, prediction, encoding, and decoding on the current layer, thereby reducing the time complexity of attribute transformation encoding and decoding, and will not have any impact on the attribute coding efficiency.
  • the decoding method proposed in the embodiment of the present application when performing RAHT decoding on the attribute, determines whether the current RAHT transformation layer belongs to a skip coding layer by judging whether the number of nodes of the current layer is consistent with the number of child nodes of the current layer at each RAHT transformation layer, that is, determines whether to skip the transformation, prediction and other processes of the current layer. If the number of nodes of the current layer is consistent with the number of child nodes of the current layer, it is considered that the current layer belongs to a skip coding layer, and does not need to perform transformation, prediction, encoding and decoding processes, thereby reducing the time complexity of attribute transformation encoding ⁇ decoding, and will not have any impact on the coding efficiency of the attribute.
  • the decoding of voxel-level node attributes can also be skipped.
  • the time complexity of point cloud attribute encoding and decoding can be reduced, and the bit rate can also be saved, thereby improving the encoding and decoding performance of the point cloud.
  • FIG. 44 a schematic flow chart of an encoding method provided by an embodiment of the present application is shown. As shown in FIG. 44 , the method may include:
  • S4401 Determine a first number of voxel nodes of a current unit and a second number of reconstruction nodes of the current unit.
  • the encoding method of the embodiment of the present application is applied to a point cloud encoder (which may be referred to as "encoder" for short).
  • the method may refer to a point cloud encoding method, specifically a point cloud attribute encoding method, and more specifically, a skip encoding method for point cloud attribute RAHT transform prediction.
  • the optimization is mainly to determine whether the nodes of the current layer are skipped and whether the voxel nodes are skipped when divided into voxel levels, thereby reducing the time complexity of attribute transformation encoding and decoding, and will not have any impact on the encoding and decoding efficiency of the attribute transformation.
  • the current unit may be a current coding unit to be encoded, which may be a slice.
  • the method may further include: dividing the nodes in the current unit to determine at least one layer; when the nodes of the last layer are divided into voxel levels, determining the voxel nodes of the current unit and the first number of voxel nodes.
  • the order of RAHT attribute transformation is to divide from the root node in sequence until it is divided into the voxel level, specifically, the division is stopped when it is divided into a unit cube of size 1 ⁇ 1 ⁇ 1, thereby completing the encoding and reconstruction of the entire point cloud attribute.
  • the layer obtained by downsampling along the Z direction, Y direction and X direction each time is a RAHT transformation layer, that is, layer. Then until it is divided into a unit cube of size 1 ⁇ 1 ⁇ 1, it means that it has been divided into the voxel level, at this time the voxel node and the first number of voxel nodes can be determined.
  • the voxel node represents the node corresponding to the division from the root node to the voxel level, and the size of the voxel node is 1 ⁇ 1 ⁇ 1;
  • the reconstruction node represents the node in the current unit that performs attribute reconstruction.
  • the geometric information of the nodes in the current unit has been completely encoded.
  • the second number of reconstructed nodes in the current unit can be determined, so that it can be determined whether the first number is consistent with the second number.
  • S4402 Determine whether the attribute information of the voxel node of the current unit is skipped for encoding based on the first quantity and the second quantity.
  • the first number is the same as the second number, encoding the attribute information of the voxel node of the current unit is skipped.
  • attribute encoding is performed on the repeated nodes in the voxel nodes, and encoding of attribute information of the remaining voxel nodes except the repeated nodes in the voxel nodes is skipped.
  • duplicate nodes there may be duplicate nodes in the voxel nodes of the current unit division, so the first number and the second number may be inconsistent. If there are no duplicate nodes in the voxel nodes of the current unit division, the first number and the second number are consistent; conversely, if there are duplicate nodes in the voxel nodes of the current unit division, the first number and the second number are inconsistent, and it is necessary to perform attribute encoding on the duplicate nodes.
  • duplicate nodes also referred to as "duplicate points” refer to multiple nodes with the same geometric information but different attribute information.
  • performing attribute encoding on repeated nodes in voxel nodes may include: performing encoding processing on attribute information of the repeated nodes, and writing the obtained encoding bits into a bitstream.
  • when encoding the attribute information of a repeated node it may include: setting a timer; when the attribute information of the repeated node begins to be encoded, starting the timer, and when the timer reaches a preset value, determining that all the attribute information of the repeated node has been encoded.
  • the third number of repeated nodes in the voxel node can also be determined based on the difference between the first number and the second number; wherein the setting of the timer is correlated with the third number.
  • a timer may be used to determine whether all attribute reconstruction values of repeated nodes have been encoded. If the timer is in a positive counting mode, then when the attribute reconstruction value of the repeated node begins to be encoded, the initial value of the timer is 0, and when the timer counts to a third number, all attribute reconstruction values of the repeated node are now completely encoded; if the timer is in a countdown mode, then when the attribute reconstruction value of the repeated node begins to be encoded, the initial value of the timer is the third number, and when the timer counts to 0, all attribute reconstruction values of the repeated node are now completely encoded. After all attribute reconstruction values of repeated nodes are completely encoded, it means that there are no repeated nodes in subsequent voxel nodes, and attribute encoding of these nodes can be skipped.
  • the method may further include:
  • S4501 Determine the attribute reconstruction value of the voxel node of the current unit according to the first quantity and the second quantity.
  • determining the attribute reconstruction value of the voxel node of the current unit according to the first number and the second number may include: setting the attribute reconstruction value of the voxel node of the current unit to the attribute reconstruction value of the reconstruction node of the current unit.
  • encoding the attribute information of the voxel node of the current unit may be skipped, and the attribute reconstruction value of the voxel node may be directly copied as the attribute reconstruction value of the reconstruction node of the current unit.
  • determining the attribute reconstruction value of the voxel node of the current unit based on the first number and the second number may include: using the attribute reconstruction value of the repeated node as the attribute reconstruction value of the first reconstruction node in the current unit, and setting the attribute reconstruction value of the remaining voxel nodes in the voxel node except the repeated node as the attribute reconstruction value of the remaining reconstruction nodes in the current unit except the first reconstruction node.
  • the attribute reconstruction value of the duplicate node can be determined, and then the attribute reconstruction value of the duplicate node is used as the attribute reconstruction value of the first reconstruction node in the current unit, and the attribute reconstruction value of the remaining voxel nodes other than the duplicate node is copied as the attribute reconstruction value of the remaining reconstruction nodes other than the first reconstruction node in the current unit.
  • the attribute information of the duplicate nodes needs to be encoded so that the decoding end can decode and determine the attribute reconstruction values of these duplicate nodes.
  • FIG. 46 is a flow chart of another encoding method provided in the embodiment of the present application. As shown in FIG. 46, the method may include:
  • S4601 Determine a fourth number of nodes in the current layer and a fifth number of child nodes corresponding to the nodes in the current layer; wherein the fourth number and a fifth quantity are used to determine whether to skip encoding the current layer.
  • the fourth number of nodes in the current layer can be determined first, and the fifth number of child nodes corresponding to the nodes in the current layer can be determined at the same time.
  • the number of nodes in the current layer i.e., the "fourth number”
  • the number of child nodes of the nodes in the current layer i.e., the "fifth number”
  • the number of nodes in the current layer i.e., the "fifth number
  • the number of child nodes of the nodes in the current layer i.e., the "fifth number
  • a RAHT attribute transformation structure based on the geometric information of the points in the point cloud, and the encoding can be performed in the order from the root node to the child node.
  • the geometric information of the current layer node is used to restore the child nodes of the current layer in the order of Z, Y and X, and then the attributes of the nodes of the previous layer are used to predictively encode the attributes of the nodes of the current layer, so as to restore the attribute reconstruction value of the nodes of the current layer until the transformation is to the voxel level, and then a RAHT attribute transformation structure including at least one RAHT transformation layer can be obtained.
  • the RAHT attribute transformation can be performed based on the order of the octree hierarchy.
  • the voxel level can be continuously transformed until the root node, thereby constructing the octree.
  • the attribute prediction transformation encoding is also performed based on the hierarchical order of the octree, but the root node is continuously transformed until the voxel level.
  • a layer obtained by downsampling in a preset direction is a RAHT transformation layer, such as the current layer.
  • the current layer may include at least one point.
  • the at least one point in the current layer when encoding the current layer, it can be used as a node to be encoded in the current layer.
  • each point in the current layer corresponds to a geometric information and an attribute information; wherein the geometric information represents the spatial relationship of the point, and the attribute information represents the relevant information of the attribute of the point.
  • the attribute information may be color information, or reflectivity or other attributes, which is not specifically limited in the embodiments of the present application.
  • the attribute information may be color information in any color space.
  • the attribute information may be color information in an RGB space, or may be color information in a YUV space, or may be color information in a YCbCr space, etc., which is not specifically limited in the embodiments of the present application.
  • the non-voxel level node attribute transformation and inverse transformation are completed first, and then the voxel level node transformation is completed, because there may be a situation in the point cloud, that is, there are duplicate nodes in the point cloud.
  • the fourth number can represent the number of occupied nodes of the current layer; and the fifth number can represent the number of nodes to be encoded.
  • the fourth number is the number of valid nodes (i.e., occupied nodes) of the current layer, and for non-voxel level nodes, the fifth number is the number of valid child nodes (i.e., occupied child nodes) of the nodes of the current layer, and for voxel level nodes, the fifth number is the number of nodes to be encoded.
  • the geometric information of the nodes of the current layer may be determined first; and then the child nodes corresponding to the nodes of the current layer and the fifth quantity may be determined based on the geometric information.
  • the current node of the current layer when using the geometric information of the current node to determine the corresponding child node, you can choose to use the geometric information of the current node for upsampling to obtain the child nodes occupied by the current node (the number of child nodes is N, where the maximum value of N is 8).
  • the number of nodes of the current layer when encoding and decoding the attribute information of the nodes of the current layer, the number of nodes of the current layer, that is, the fourth number, can be obtained first; at the same time, after the child nodes of the nodes of the current layer are restored using the geometric information of the nodes of the current layer, the number of child nodes of the nodes of the current layer, that is, the fifth number, can be obtained.
  • the attribute reconstruction value of the child node corresponding to the node of the current layer can be further determined based on the fourth quantity and the fifth quantity.
  • the fourth number and the fifth number can be used to encode and decode the attribute information of whether to skip the coding layer (current layer).
  • the RAHT transform is only valid for nodes with neighboring points, if the number of nodes in the current layer is exactly the same as the number of child nodes of the nodes in the current layer, it can be indicated that each node in the current layer has only one child node; in this case, the current layer will not generate AC coefficients (high-frequency coefficients), so it is possible to choose to skip the transformation, prediction, and other processes performed on the nodes of the current layer in sequence.
  • the current layer can be adaptively determined. Whether the layer can skip encoding and decoding.
  • the key to determining whether to skip encoding and decoding processing of the current layer is whether the number of nodes and the number of child nodes of the current layer are the same, that is, whether the fourth number and the fifth number are the same.
  • the method may further include: if the fourth quantity and the fifth quantity are the same, determining the attribute reconstruction value of the node of the current layer as the attribute reconstruction value of the child node corresponding to the node of the current layer.
  • the fourth number and the fifth number corresponding to the nodes of the current layer are the same, it can be determined that the number of nodes in the current layer is the same as the number of child nodes corresponding to the nodes of the current layer. Then, it can be considered that for each node in the current layer, there is only one corresponding child node.
  • the RAHT transform is only valid for nodes with neighboring points, during the RAHT attribute transformation process, if each node in the current layer corresponds to only one child node, then it can be considered that the current layer will not generate AC coefficients. Therefore, it can be chosen not to transform, predict, etc. the nodes of the current layer in sequence, that is, skip processing the current layer. At this time, it can be called "skipping the coding layer.”
  • the fourth quantity and the fifth quantity corresponding to the nodes of the current layer are the same, that is, it is determined that the current layer is a skip coding layer, then it is possible to choose to skip the transformation, prediction, and other processes of the nodes of the current layer in sequence, and instead directly determine the attribute reconstruction value of the node of the current layer as the attribute reconstruction value of the child node corresponding to the node of the current layer.
  • each node of the current layer corresponds to only one child node, then it is possible to choose not to perform transformation, prediction, and other processes on the nodes of the current layer in sequence, thereby reducing the complexity of RAHT attribute transformation encoding and decoding.
  • the fourth number and the fifth number are the same, and the current layer is determined to be a skip coding layer, then you can choose to skip the current layer to the next layer, and then use the next layer as the current layer, and continue to determine whether to skip coding the nodes of the next layer.
  • the sixth number of the next layer of child nodes corresponding to the child node can be determined first; then, based on the fifth number and the sixth number, the attribute reconstruction value of the next layer of child nodes corresponding to the child node can be determined.
  • the sixth number can be the number of valid child nodes in the next layer (i.e., the occupied child nodes in the next layer) of the child node corresponding to the node of the current layer; if the child node corresponding to the node of the current layer is a voxel-level node, then the sixth number can be the number of nodes to be encoded.
  • the fifth number and the sixth number can be used to encode and decode the attribute information of whether to skip the coding layer (the next layer of the current layer).
  • the fifth number and the sixth number are the same, then it may be possible to choose not to perform transformation, prediction, etc. on the child nodes of the node of the current layer in sequence, that is, skip processing the child nodes of the current layer, and instead directly determine the attribute reconstruction value of the child node corresponding to the node of the current layer as the attribute reconstruction value of the next layer of child nodes of the child node corresponding to the node of the current layer.
  • the method may also include: if the fourth quantity and the fifth quantity corresponding to the node of the current layer are different, determining the attribute prediction value of the child node corresponding to the node of the current layer according to the node of the current layer; performing a RAHT transform based on the attribute prediction value of the child node to determine the reconstruction value of the high-frequency coefficient and the low-frequency coefficient corresponding to the node of the current layer; performing a RAHT inverse transform based on the reconstruction value of the high-frequency coefficient and the low-frequency coefficient to determine the attribute reconstruction value of the child node.
  • the current layer will still generate AC coefficients, so the nodes of the current layer can continue to be transformed, predicted, etc. in sequence without skipping the processing of the current layer.
  • the current layer can be used as a non-skipped coding layer.
  • when determining the attribute prediction value of the child node corresponding to the node of the current layer based on the node of the current layer it can include: determining the adjacent nodes corresponding to the node of the current layer; and determining the attribute prediction value of the child node corresponding to the node of the current layer based on the attribute reconstruction value and relative distance parameter corresponding to the adjacent nodes.
  • the adjacent node may refer to the neighboring node of the current node.
  • the relative distance parameter corresponding to the adjacent node may represent the spatial geometric distance between the child node corresponding to the node of the current layer and the corresponding adjacent node.
  • the current node includes two sub-nodes, sub-node 1 and sub-node 2, and the relative distance parameter between the current node and the adjacent node may include the spatial geometric distance between sub-node 1 and the adjacent node, and may also include the spatial geometric distance between sub-node 2 and the adjacent node.
  • the reconstructed attributes (attribute reconstruction values) of the neighboring nodes of the current node and the spatial geometric distance of each neighboring node from the child node of the current node can be used to perform linear fitting, and finally obtain the attribute prediction value of each child node of the current node.
  • the 19 adjacent nodes of the current node can be determined first. Point, then use the spatial geometric distance between the adjacent node and each child node of the current node to perform linear weighted prediction on the attributes of each child node, and finally obtain the attribute prediction value of each child node.
  • when performing a RAHT transform based on the attribute prediction value of the child node to determine the reconstructed value of the high-frequency coefficient and the low-frequency coefficient corresponding to the node of the current layer it can include: performing a RAHT transform based on the attribute prediction value of the child node to determine the predicted value of the high-frequency coefficient and the low-frequency coefficient corresponding to the node of the current layer; performing a RAHT transform based on the attribute value of the child node to determine the high-frequency coefficient and the low-frequency coefficient corresponding to the node of the current layer; determining the reconstructed value of the high-frequency coefficient corresponding to the node of the current layer according to the predicted value of the high-frequency coefficient corresponding to the node of the current layer and the high-frequency coefficient corresponding to the node of the current layer.
  • when determining the reconstructed values of the high-frequency coefficients corresponding to the nodes of the current layer based on the predicted values of the high-frequency coefficients corresponding to the nodes of the current layer and the high-frequency coefficients corresponding to the nodes of the current layer it can include: determining the coefficient residuals corresponding to the nodes of the current layer based on the predicted values of the high-frequency coefficients corresponding to the nodes of the current layer and the high-frequency coefficients corresponding to the nodes of the current layer; determining the reconstructed values of the high-frequency coefficients corresponding to the nodes of the current layer based on the coefficient residuals corresponding to the nodes of the current layer.
  • the determining the reconstructed values of the high-frequency coefficients corresponding to the nodes of the current layer based on the coefficient residuals corresponding to the nodes of the current layer it can include: dequantizing the quantized coefficient residuals to determine the dequantized residual values corresponding to the nodes of the current layer; and determining the reconstructed values of the high-frequency coefficients corresponding to the nodes of the current layer based on the dequantized residual values corresponding to the nodes of the current layer and the predicted values of the high-frequency coefficients corresponding to the nodes of the current layer.
  • the attribute prediction value of the corresponding child node can be used to perform RAHT attribute transformation, so as to obtain the corresponding DC coefficient and AC coefficient.
  • the DC coefficient is the low-frequency coefficient
  • the AC coefficient is the high-frequency coefficient.
  • the AC coefficient obtained by performing RAHT attribute transformation using the attribute prediction value corresponding to the child node can be understood as the predicted value of the AC coefficient corresponding to the current node.
  • the attribute value of the child node can also be used to perform a RAHT transformation to determine the AC coefficient and DC coefficient corresponding to the node of the current layer.
  • the AC coefficient obtained by performing a RAHT attribute transformation using the attribute value corresponding to the child node can be understood as the original value of the AC coefficient corresponding to the current node.
  • the coefficient residual corresponding to the node of the current layer can be determined; then the quantized coefficient residual is inversely quantized to determine the inversely quantized residual value corresponding to the node of the current layer; then based on the inversely quantized residual value corresponding to the node of the current layer and the predicted value of the AC coefficient corresponding to the node of the current layer, the reconstructed value of the AC coefficient corresponding to the node of the current layer can be determined.
  • the inverse quantization residual value corresponding to the node of the current layer and the predicted value of the AC coefficient corresponding to the node of the current layer can be summed up to obtain the reconstructed value of the AC coefficient corresponding to the node of the current layer.
  • the method may also include: quantizing the coefficient residuals to determine the quantized coefficient residuals corresponding to the nodes of the current layer; encoding the quantized coefficient residuals and writing the obtained coded bits into the bitstream.
  • the quantized coefficient residual is written into the bit stream, and the quantized coefficient residual can be obtained by decoding the bit stream at the decoding end. Then, based on the predicted value of the high-frequency coefficient and the quantized coefficient residual, the reconstructed value of the high-frequency coefficient corresponding to the node of the current layer can be determined.
  • g′ L,2x,y,z and g′ L,2x+1,y,z are two attribute DC coefficients of neighboring points in the L layer.
  • the information of the L-1 layer is the AC coefficient f′ L-1,x,y,z and the DC coefficient g′ L-1,x,y,z ; then, f′ L-1,x,y,z will no longer be transformed and will be directly quantized and encoded, and g′ L-1,x,y,z will continue to look for neighbors for transformation.
  • the weights (the number of non-empty child nodes in the node) corresponding to g′ L,2x,y,z and g′ L,2x+2,y ,z are w′ L,2x,y,z and w′ L,2x+1,y,z (abbreviated as w′ 0 and w′ 1 ) respectively, and the weight of g′ L-1,x,y,z is w′ L-1,x,y,z .
  • the general transformation formula is:
  • the inverse transformation of RAHT is performed according to the DC coefficient and AC coefficient of the child node of the current node, and the attribute reconstruction value of the child node of the current node can be restored.
  • the inverse transformation of RAHT (also referred to as "RAHT inverse transformation” or "RAHT inverse transformation”) is shown in the aforementioned FIG. 35B.
  • the nodes of the current layer can continue to be transformed, predicted, etc. in sequence.
  • the reconstruction attributes of the adjacent nodes of the point and the spatial geometric distance of each adjacent node from each child node of the current node are linearly fitted to obtain the predicted attributes of each child node of the current node; then, the predicted attributes of each child node are used to perform RAHT attribute transformation to obtain the corresponding DC and AC coefficients.
  • the attributes of each child node of the current node can be transformed by RAHT attribute transformation to obtain DC and AC coefficients; then, the predicted value of the AC coefficient obtained by the prediction node can be used to predict the AC coefficient of the current node, thereby obtaining the AC prediction residual coefficient (coefficient residual) of each child node, and then the coefficient residual can be quantized and encoded.
  • the inverse quantization residual value of the AC prediction residual coefficient and the predicted value of the AC coefficient can also be used to restore the AC reconstruction coefficient of the current node (the reconstruction value of the high-frequency coefficient), and finally the AC coefficient and DC coefficient of the current node are used to perform RAHT inverse transformation to restore the attribute reconstruction value of each child node of the current node.
  • the nodes of the current layer can be transformed, predicted, and the like in sequence to determine the attribute reconstruction values of the child nodes corresponding to the nodes of the current layer; then, for the child nodes of the current layer nodes, the sixth number of the next-layer child nodes corresponding to the child nodes can be determined first; and then, based on the fifth number and the sixth number, the attribute reconstruction values of the next-layer child nodes corresponding to the child nodes can be determined.
  • step S4601 to step S4602 is continuously repeated, starting from the root node of the RAHT transformation until the last node of the leaf node layer of the RAHT, thereby completing the attribute encoding of the entire RAHT transformation.
  • the method may also include: determining prediction mode identification information; when the prediction mode identification information indicates that the current unit starts the skip coding mode, executing the first quantity and the second quantity determination steps, and/or executing the fourth quantity and the fifth quantity determination steps.
  • the prediction mode identification information is at least one of the following high-level syntax elements: a syntax element corresponding to an attribute parameter set (Attribute Parameter Set, APS) and a syntax element corresponding to an attribute block header information (Attribute Block Head, ABH).
  • the prediction mode identification information may also be encoded, and the obtained encoding bits may be written into the bitstream. If the current unit starts the skip coding mode, the value of the prediction mode identification information is determined to be the first value; if the current unit does not start the skip coding mode, the value of the prediction mode identification information is determined to be the second value. The following describes whether the voxel node of the current unit starts the skip coding mode and whether the node of the current layer starts the skip coding mode.
  • the method may further include: determining first prediction mode identification information; and executing the first quantity and the second quantity determination steps when the first prediction mode identification information indicates that the voxel node of the current unit starts the skip coding mode.
  • the first prediction mode identification information may also be encoded, and the obtained encoding bits may be written into the bitstream. If the voxel node of the current unit starts the skip encoding mode, the value of the first prediction mode identification information is determined to be the first value; if the voxel node of the current unit does not start the skip encoding mode, the value of the first prediction mode identification information is determined to be the second value.
  • the first number and the second number can be further determined at this time, and then the voxel node of the current unit is determined to skip coding based on the size of the first number and the second number. Specifically, if the two are consistent, the node attributes at the voxel level are skipped, otherwise, a timer is used.
  • the number of remaining repeated nodes of the timer is zero, it means that there are no repeated nodes in the subsequent points, and the attribute coding of the subsequent points can also be skipped, and the attribute reconstruction value of the subsequent voxel node is directly copied to the attribute reconstruction value of the final remaining reconstruction point.
  • the method may further include: determining second prediction mode identification information; and when the second prediction mode identification information indicates that a node of the current layer starts a skip coding mode, executing the steps of determining the fourth quantity and the fifth quantity.
  • the second prediction mode identification information may also be encoded, and the obtained encoding bits may be written into the bitstream. If the node of the current layer starts the skip coding mode, the value of the second prediction mode identification information is determined to be the first value; if the node of the current layer does not start the skip coding mode, the value of the second prediction mode identification information is determined to be the second value.
  • the fourth number and the fifth number can be further determined at this time, and then the size of the fourth number and the fifth number is used to determine whether the node of the current layer skips coding, that is, whether the current layer is a skip coding layer. Specifically, if the two are consistent, the current layer is determined to be a skip coding layer, and it is no longer necessary to encode the attribute reconstruction value of the child node corresponding to the node of the current layer; if the two are inconsistent, it is determined that the current layer does not belong to the skip coding layer, and RAHT prediction and encoding can be performed according to the encoding method of the relevant technology.
  • the first value is different from the second value, and the first value and the second value can be in parameter form or in digital form.
  • the first prediction mode identification information and the second prediction mode identification information can be parameters written in the profile, or can be the value of a flag, which is not specifically limited here.
  • the first value can be set to 1 and the second value can be set to 0; or, the first value can be set to 0 and the second value can be set to 1; or, the first value can be set to true and the second value can be set to false; or, the first value can be set to false and the second value can be set to
  • the first value is set to 1 and the second value is set to 0, but this is not specifically limited.
  • the node of the current layer starts the skip coding mode, then it can be determined that the value of the second prediction mode identification information is 1, and then the fourth quantity and the fifth quantity corresponding to the node of the current layer can be further determined according to the above method; if the node of the current layer does not start the skip coding mode, then it can be determined that the value of the second prediction mode identification information is 0, and the node of the current layer can be attribute encoded in the common intra-frame prediction method or inter-frame prediction method.
  • the first quantity and the second quantity determination steps can be executed, that is, the encoding process shown in Figure 44 is executed; if the value of the second prediction mode identification information written in the code stream is the first value, that is, it is determined that the node of the current layer starts the skip coding mode, then the fourth quantity and the fifth quantity determination process can be executed, that is, the encoding process shown in Figure 46 is executed.
  • an embodiment of the present application also provides a code stream, which is generated by bit encoding based on information to be encoded; wherein the information to be encoded may include at least one of the following: prediction mode identification information, attribute information of repeated nodes of the current unit, and quantized coefficient residuals corresponding to nodes of the current layer; wherein the prediction mode identification information is used to indicate whether the current unit starts the skip coding mode.
  • the information to be encoded can be written into the bit stream; then the encoding end transmits it to the decoding end, and subsequently at the decoding end, this information, such as the prediction mode identification information, can be obtained by decoding the bit stream, and then it can be determined whether the current unit starts the skip coding mode.
  • the current layer when encoding attribute information, if the number of nodes in the current layer is consistent with the number of child nodes in the current layer, the current layer is considered to be a skip coding layer, and therefore there is no need to perform transformation, prediction, encoding, and decoding on the current layer, thereby reducing the time complexity of attribute transformation encoding and decoding, and will not have any impact on the attribute encoding efficiency.
  • the present embodiment provides a coding method, in which a coding method for skipping the current layer is proposed.
  • a coding method for skipping the current layer is proposed.
  • the time complexity of coding and decoding can be reduced under the premise of ensuring the coding and decoding efficiency remains unchanged.
  • a coding method for skipping voxel-level nodes is also proposed. First, the number of voxel nodes and the number of reconstructed nodes can be obtained.
  • the number of reconstructed nodes is consistent with the number of voxel nodes, it is considered that there are no duplicate nodes in the current unit, and the encoding of voxel-level node attributes can also be skipped.
  • the time complexity of point cloud attribute coding and decoding can be reduced, and the bit rate can also be saved, thereby improving the coding and decoding performance of the point cloud.
  • the embodiment of the present application first defines a RAHT attribute coding layer.
  • the current attribute RAHT transform coding order is to divide from the root node in sequence until it is divided into the voxel level (1 ⁇ 1 ⁇ 1), thereby completing the attribute encoding and attribute reconstruction of the entire point cloud.
  • the layer obtained by downsampling once along the Z direction, Y direction, and X direction is a RAHT transform layer, that is, layer. Specifically shown in Figure 42.
  • an algorithm for skipping coding layers is introduced.
  • the number of nodes in the current layer can be obtained.
  • the number of child nodes of the current layer nodes can be obtained. Based on the size of the two, it is determined whether the current layer belongs to the skip coding layer:
  • the current layer belongs to the skip coding layer
  • the current layer is a non-skipped coding layer.
  • Step 1 Determine whether the number of nodes in the current layer is consistent with the number of child nodes in the current layer. If they are consistent, the current layer belongs to a skip coding layer; otherwise, it is a non-skipped coding layer;
  • Step 2 If the current layer does not belong to a skip coding layer, transform, predict and encode according to the encoding method of the related technology.
  • Step 3 If the current layer belongs to a skip coding layer, directly skip the current layer to the next layer.
  • Step 4 Repeat the above steps until encoding is done at the voxel level.
  • Step 5 For voxel-level nodes, before encoding the attributes of the voxel-level nodes, first determine whether the number of voxel-level nodes in the current coding unit is consistent with the number of reconstructed nodes. If they are consistent, skip encoding the attributes of the voxel-level points; otherwise, a timer is used. When the number of remaining duplicate points of the timer is zero, it means that there are no duplicate points in the subsequent nodes, and the attribute encoding of the subsequent nodes can also be skipped.
  • Step 1 Determine whether the number of nodes in the current layer is consistent with the number of child nodes in the current layer. If they are consistent, the current layer belongs to a skip decoding layer; otherwise, it is a non-skip decoding layer;
  • Step 2 If the current layer does not belong to a skip decoding layer, prediction and decoding are performed according to the decoding method of the related art.
  • Step 3 If the current layer belongs to a skip decoding layer, directly skip the current layer to the next layer.
  • Step 4 Repeat the above steps until decoding is completed at the voxel level.
  • Step 5 For voxel-level nodes, before decoding the attributes of the voxel-level nodes, first determine whether the number of voxel-level nodes in the current decoding unit is consistent with the number of reconstructed nodes. If they are consistent, skip decoding the attributes of the voxel-level points; otherwise, a timer is used. When the number of remaining duplicate points of the timer is zero, it means that there are no duplicate points in the subsequent nodes. Similarly, the attribute decoding of the subsequent nodes can be skipped, and the attribute reconstruction values of the subsequent remaining voxel nodes are directly copied as the attribute reconstruction values of the final remaining reconstructed nodes.
  • a coding method for skipping coding layers is proposed here, by using the number of nodes and the number of sub-nodes of the current layer to adaptively determine whether the current layer can skip coding, so that the time complexity of coding and decoding can be reduced under the premise of ensuring that the coding and decoding efficiency remains unchanged.
  • the codec end first completes the encoding and decoding of non-node layer attribute information (that is, the size of the node is greater than or equal to 1 ⁇ 1 ⁇ 1), and finally encodes and decodes the attribute information of the voxel-level nodes.
  • the reason is that there will be duplicate points in the point cloud, so it is necessary to first complete the encoding and decoding of the attribute information of the non-voxel-level points, and then complete the encoding and decoding of the voxel-level point attribute information.
  • the number of voxel-level reconstruction points can be obtained first.
  • the number of reconstruction points is consistent with the number of nodes that need to be reconstructed, it is considered that there are no duplicate points in the current coding unit (such as a slice), and the encoding/decoding of the voxel-level point attribute information can also be skipped.
  • the current layer belongs to a skip coding layer, that is, no transformation, prediction, encoding and decoding processes are required, thereby reducing the time complexity of attribute transformation encoding/decoding, and will not have any impact on the coding efficiency of the attribute.
  • the encoder 470 may include: a first determination unit 4701 and an encoding unit 4702, wherein:
  • a first determining unit 4701 is configured to determine a first number of voxel nodes of a current unit and a second number of reconstruction nodes of the current unit;
  • the encoding unit 4702 is configured to determine whether the attribute information of the voxel node of the current unit is to be skipped for encoding according to the first quantity and the second quantity.
  • the encoding unit 4702 is further configured to skip encoding the attribute information of the voxel node if the first number is the same as the second number; if the first number is different from the second number, attribute encoding is performed on the repeated nodes in the voxel node, and the attribute information of the remaining voxel nodes except the repeated nodes in the voxel node is skipped.
  • the encoding unit 4702 is further configured to encode the attribute information of the repeated nodes and write the obtained encoded bits into the bit stream.
  • the encoding unit 4702 is further configured to set a timer; when the attribute information of the repeated node begins to be encoded, the timer is started, and when the timer reaches a preset value, it is determined that all the attribute information of the repeated node has been encoded.
  • the first determination unit 4701 is further configured to determine a third number of repeated nodes in the voxel node based on a difference between the first number and the second number; wherein the setting of the timer is associated with the third number.
  • the encoder 470 may further include a first reconstruction unit 4703 configured to determine a property reconstruction value of a voxel node of a current unit according to the first quantity and the second quantity.
  • the first reconstruction unit 4703 is further configured to set the attribute reconstruction value of the voxel node of the current unit to the attribute reconstruction value of the reconstruction node of the current unit when the first number is the same as the second number.
  • the first reconstruction unit 4703 is also configured to use the attribute reconstruction value of the repeated node as the attribute reconstruction value of the first reconstruction node in the current unit when the first number is different from the second number, and to set the attribute reconstruction value of the remaining voxel nodes in the voxel node except the repeated node as the attribute reconstruction value of the remaining reconstruction nodes in the current unit except the first reconstruction node.
  • the first determination unit 4701 is further configured to divide the nodes in the current unit to determine at least one layer; and when the nodes of the last layer are divided into voxel levels, determine the voxel nodes of the current unit and the first number of voxel nodes.
  • the at least one layer includes a current layer
  • the first determining unit 4701 is further configured to determine a fourth number of nodes of the current layer and a fifth number of child nodes corresponding to the nodes of the current layer; wherein the fourth number and the fifth number are used to determine whether to skip encoding the current layer;
  • the first reconstruction unit 4703 is further configured to determine the attribute reconstruction value of the child node corresponding to the node of the current layer according to the fourth quantity and the fifth quantity.
  • the fourth number represents the number of occupied nodes in the current layer
  • the fifth number represents the number of occupied child nodes or the number of nodes to be encoded in the nodes of the current layer.
  • the first determination unit 4701 is further configured to determine the attribute reconstruction value of the node of the current layer as the attribute reconstruction value of the child node corresponding to the node of the current layer if the fourth number and the fifth number are the same.
  • the first determining unit 4701 is further configured to determine a sixth number of next-layer child nodes corresponding to the child node if the fourth number is the same as the fifth number;
  • the first reconstruction unit 4703 is further configured to determine the attribute reconstruction value of the next layer of child nodes corresponding to the child node according to the fifth number and the sixth number.
  • the first reconstruction unit 4703 is further configured to determine the attribute prediction value of the child node corresponding to the node of the current layer according to the node of the current layer if the fourth quantity and the fifth quantity are different; perform RAHT transformation based on the attribute prediction value of the child node and the attribute value of the child node respectively to determine the reconstruction value of the high-frequency coefficient and the low-frequency coefficient corresponding to the node of the current layer; and perform RAHT inverse transformation based on the reconstruction value of the high-frequency coefficient and the low-frequency coefficient to determine the attribute reconstruction value of the child node.
  • the first determination unit 4701 is further configured to determine the adjacent nodes corresponding to the nodes of the current layer; and determine the attribute prediction values of the child nodes corresponding to the nodes of the current layer based on the attribute reconstruction values and relative distance parameters corresponding to the adjacent nodes.
  • the first reconstruction unit 4703 is also configured to perform a RAHT transform based on the attribute prediction value of the child node to determine the prediction value of the high-frequency coefficient and the low-frequency coefficient corresponding to the node of the current layer; perform a RAHT transform based on the attribute value of the child node to determine the high-frequency coefficient and the low-frequency coefficient corresponding to the node of the current layer; and determine the reconstructed value of the high-frequency coefficient corresponding to the node of the current layer based on the prediction value of the high-frequency coefficient corresponding to the node of the current layer and the high-frequency coefficient corresponding to the node of the current layer.
  • the first reconstruction unit 4703 is also configured to determine the coefficient residuals corresponding to the nodes of the current layer based on the predicted values of the high-frequency coefficients corresponding to the nodes of the current layer and the high-frequency coefficients corresponding to the nodes of the current layer; and determine the reconstructed values of the high-frequency coefficients corresponding to the nodes of the current layer based on the coefficient residuals corresponding to the nodes of the current layer.
  • the first determining unit 4701 is further configured to quantize the coefficient residual to determine the quantized coefficient residual corresponding to the node of the current layer;
  • the encoding unit 4702 is further configured to perform encoding processing on the quantized coefficient residual and write the obtained encoding bits into the bit stream.
  • the first reconstruction unit 4703 is also configured to dequantize the quantized coefficient residual to determine the dequantized residual value corresponding to the node of the current layer; and determine the reconstructed value of the high-frequency coefficient corresponding to the node of the current layer based on the dequantized residual value corresponding to the node of the current layer and the predicted value of the high-frequency coefficient corresponding to the node of the current layer.
  • the first determination unit 4701 is further configured to determine geometric information of the nodes of the current layer; determine the child nodes corresponding to the nodes of the current layer and the fifth quantity according to the geometric information.
  • the first determining unit 4701 is further configured to determine prediction mode identification information
  • the encoding unit 4702 is further configured to, when the prediction mode identification information indicates that the current unit starts the skip encoding mode, perform the steps of determining the first quantity and the second quantity, and/or perform the steps of determining the fourth quantity and the fifth quantity.
  • the prediction mode identification information is at least one of the following high-level syntax elements: a syntax element corresponding to the attribute parameter set and a syntax element corresponding to the attribute block header information.
  • the encoding unit 4702 is further configured to encode the prediction mode identification information and write the obtained encoding bits into the bitstream.
  • a "unit” may be a part of a circuit, a part of a processor, a part of a program or software, etc., and of course, it may be a module, or it may be non-modular.
  • the components in the present embodiment may be integrated into a processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.
  • the above-mentioned integrated unit may be implemented in the form of hardware or in the form of a software functional module.
  • the integrated unit is implemented in the form of a software function module and is not sold or used as an independent product, it can be stored in a computer-readable storage medium.
  • the technical solution of this embodiment is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product.
  • the computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, server, or network device, etc.) or a processor to perform all or part of the steps of the method described in this embodiment.
  • the aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk, etc., various media that can store program codes.
  • an embodiment of the present application provides a computer-readable storage medium, which is applied to the encoder 470.
  • the computer-readable storage medium stores a computer program, and when the computer program is executed by the first processor, the method described in any one of the aforementioned embodiments is implemented.
  • the encoder 470 may include: a first communication interface 4801, a first memory 4802 and a first processor 4803; each component is coupled together through a first bus system 4804. It can be understood that the first bus system 4804 is used to realize the connection and communication between these components.
  • the first bus system 4804 also includes a power bus, a control bus and a status signal bus. However, for the sake of clarity, various buses are marked as the first bus system 4804 in Figure 48. Among them:
  • the first communication interface 4801 is used for receiving and sending signals during the process of sending and receiving information between other external network elements;
  • a first memory 4802 used to store a computer program that can be run on the first processor 4803;
  • the first processor 4803 is configured to, when running the computer program, execute:
  • the first memory 4802 in the embodiment of the present application can be a volatile memory or a non-volatile memory, or can include both volatile and non-volatile memories.
  • the non-volatile memory can be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a flash memory.
  • the volatile memory can be a random access memory (RAM), which is used as an external cache.
  • RAM static RAM
  • DRAM dynamic RAM
  • SDRAM synchronous DRAM
  • DDRSDRAM double data rate synchronous DRAM
  • ESDRAM enhanced synchronous DRAM
  • SLDRAM synchronous link DRAM
  • DRRAM direct RAM bus RAM
  • the first processor 4803 may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by the hardware integrated logic circuit or software instructions in the first processor 4803.
  • the above-mentioned first processor 4803 can be a general-purpose processor, a digital signal processor (Digital Signal Processor, DSP), an application-specific integrated circuit (Application Specific Integrated Circuit, ASIC), a field programmable gate array (Field Programmable Gate Array, FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components.
  • DSP Digital Signal Processor
  • ASIC Application Specific Integrated Circuit
  • FPGA Field Programmable Gate Array
  • the methods, steps and logic block diagrams disclosed in the embodiments of the present application can be implemented or executed.
  • the general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc.
  • the steps of the method disclosed in the embodiments of the present application can be directly embodied as a hardware decoding processor to execute, or the hardware and software modules in the decoding processor can be executed.
  • the software module can be located in a mature storage medium in the field such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory or an electrically erasable programmable memory, a register, etc.
  • the storage medium is located in the first memory 4802, and the first processor 4803 reads the information in the first memory 4802 and completes the steps of the above method in combination with its hardware.
  • the processing unit can be implemented in one or more application specific integrated circuits (Application Specific Integrated Circuits, ASIC), digital signal processors (Digital Signal Processing, DSP), digital signal processing devices (DSP Device, DSPD), programmable logic devices (Programmable Logic Device, PLD), field programmable gate arrays (Field-Programmable Gate Array, FPGA), general processors, controllers, microcontrollers, microprocessors, other electronic units for performing the functions described in this application or a combination thereof.
  • ASIC Application Specific Integrated Circuits
  • DSP Digital Signal Processing
  • DSP Device digital signal processing devices
  • PLD programmable logic devices
  • FPGA field programmable gate array
  • general processors controllers, microcontrollers, microprocessors, other electronic units for performing the functions described in this application or a combination thereof.
  • the technology described in this application can be implemented by a module (such as a process, function, etc.) that performs the functions described in this application.
  • the software code can be stored in a memory and executed by a processor.
  • the memory can be implemented in the processor or outside the processor.
  • the first processor 4803 is further configured to execute the method described in any one of the aforementioned embodiments when running the computer program.
  • This embodiment provides an encoder, in which, when reconstructing the attributes of each voxel node, the judgment conditions for whether to perform attribute encoding and decoding on each voxel node are optimized. Specifically, if there are no duplicate nodes in the current unit, that is, the first number is the same as the second number, then there is no need to encode and decode the voxel nodes of the current unit. Therefore, on the basis of ensuring the encoding and decoding efficiency of the point cloud attributes, the time complexity of encoding and decoding the point cloud attributes can be reduced, and the bit rate can also be saved, thereby improving the encoding and decoding performance of the point cloud.
  • FIG49 shows a schematic diagram of the composition structure of a decoder provided by an embodiment of the present application.
  • the decoder 490 may include a second determination unit 4901 and a second reconstruction unit 4902, wherein:
  • the second determining unit 4901 is configured to determine a first number of voxel nodes of the current unit and a second number of reconstruction nodes of the current unit; wherein the first number and the second number are used to determine whether to skip decoding the voxel nodes of the current unit;
  • the second reconstruction unit 4902 is configured to determine the attribute reconstruction value of the voxel node of the current unit according to the first quantity and the second quantity.
  • the decoder 490 may further include a decoding unit 4903 configured to skip decoding the voxel nodes of the current unit if the first number is the same as the second number.
  • the second reconstruction unit 4902 is further configured to set the attribute reconstruction value of the voxel node of the current unit to the attribute reconstruction value of the reconstruction node of the current unit when the first number is the same as the second number.
  • the second reconstruction unit 4902 is further configured to decode the code stream and determine the attribute reconstruction value of the repeated node when the first number is different from the second number; use the attribute reconstruction value of the repeated node as the attribute reconstruction value of the first reconstruction node in the current unit, and set the attribute reconstruction value of the remaining voxel nodes in the voxel node except the repeated node as the attribute reconstruction value of the remaining reconstruction nodes in the current unit except the first reconstruction node.
  • the decoding unit 4903 is further configured to set a timer; when the attribute reconstruction value of the repeated node begins to be decoded, the timer is started, and when the timer reaches a preset value, it is determined that all the attribute reconstruction values of the repeated node are decoded.
  • the second determination unit 4901 is further configured to determine a third number of repeated nodes in the voxel node based on the difference between the first number and the second number; wherein the setting of the timer is associated with the third number.
  • the second determination unit 4901 is further configured to divide the nodes in the current unit to determine at least one layer; and when the nodes of the last layer are divided into voxel levels, determine the voxel nodes of the current unit and the first number of voxel nodes.
  • the at least one layer includes a current layer
  • the second determining unit 4901 is further configured to determine a fourth number of nodes of the current layer and a fifth number of child nodes corresponding to the nodes of the current layer; wherein the fourth number and the fifth number are used to determine whether to skip decoding the current layer;
  • the fourth number represents the number of occupied nodes in the current layer; the fifth number represents the number of occupied child nodes in the nodes of the current layer or the number of nodes to be decoded.
  • the second reconstruction unit 4902 is further configured to determine the attribute reconstruction value of the node of the current layer as the attribute reconstruction value of the child node corresponding to the node of the current layer if the fourth number and the fifth number are the same.
  • the second reconstruction unit 4902 is further configured to determine the sixth number of the next layer of child nodes corresponding to the child node if the fourth number is the same as the fifth number; and determine the attribute reconstruction value of the next layer of child nodes corresponding to the child node based on the fifth number and the sixth number.
  • the second reconstruction unit 4902 is further configured to determine the attribute prediction value of the child node corresponding to the node of the current layer according to the node of the current layer if the fourth number and the fifth number are different; perform a RAHT transform based on the attribute prediction value of the child node to determine the reconstruction value of the high-frequency coefficient and the low-frequency coefficient corresponding to the node of the current layer; and perform a RAHT inverse transform based on the reconstruction value of the high-frequency coefficient and the low-frequency coefficient to determine the attribute reconstruction value of the child node.
  • the second determination unit 4901 is further configured to determine the adjacent nodes corresponding to the nodes of the current layer; and determine the attribute prediction values of the child nodes corresponding to the nodes of the current layer based on the attribute reconstruction values and relative distance parameters corresponding to the adjacent nodes.
  • the second reconstruction unit 4902 is also configured to perform RAHT transformation based on the attribute prediction value of the child node to determine the prediction value and low-frequency coefficient of the high-frequency coefficient corresponding to the node of the current layer; and determine the reconstruction value of the high-frequency coefficient corresponding to the node of the current layer according to the prediction value of the high-frequency coefficient.
  • the decoding unit 4903 is further configured to decode the bitstream to determine the quantized coefficient residual corresponding to the node of the current layer;
  • the second determination unit 4901 is further configured to determine the reconstructed value of the high-frequency coefficient corresponding to the node of the current layer according to the predicted value of the high-frequency coefficient and the quantized coefficient residual.
  • the second reconstruction unit 4902 is also configured to dequantize the quantized coefficient residual to determine the dequantized residual value corresponding to the node of the current layer; and determine the reconstructed value of the high-frequency coefficient corresponding to the node of the current layer based on the dequantized residual value corresponding to the node of the current layer and the predicted value of the high-frequency coefficient corresponding to the node of the current layer.
  • the second determination unit 4901 is further configured to determine geometric information of the nodes of the current layer; and determine the child nodes corresponding to the nodes of the current layer and the fifth quantity based on the geometric information.
  • the decoding unit 4903 is further configured to decode the code stream and determine the prediction mode identification information; and when the prediction mode identification information indicates that the current unit starts the skip decoding mode, perform the first quantity and the second quantity determination steps, and/or perform the fourth quantity and the fifth quantity determination steps.
  • the prediction mode identification information is at least one of the following high-level syntax elements: a syntax element corresponding to the attribute parameter set and a syntax element corresponding to the attribute block header information.
  • a "unit" can be a part of a circuit, a part of a processor, a part of a program or software, etc., and of course it can also be a module, or it can be non-modular.
  • the components in this embodiment can be integrated into a processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.
  • the above-mentioned integrated unit can be implemented in the form of hardware or in the form of a software functional module.
  • the integrated unit is implemented in the form of a software function module and is not sold or used as an independent product, it can be stored in a computer-readable storage medium.
  • this embodiment provides a computer-readable storage medium, which is applied to the decoder 490.
  • the computer-readable storage medium stores a computer program. When the computer program is executed by the second processor, the aforementioned The method of any one of the embodiments.
  • the decoder 490 may include: a second communication interface 5001, a second memory 5002 and a second processor 5003; each component is coupled together through a second bus system 5004. It can be understood that the second bus system 5004 is used to realize the connection and communication between these components.
  • the second bus system 5004 also includes a power bus, a control bus and a status signal bus. However, for the sake of clarity, various buses are marked as the second bus system 5004 in Figure 50. Among them:
  • the second communication interface 5001 is used for receiving and sending signals during the process of sending and receiving information with other external network elements;
  • the second memory 5002 is used to store a computer program that can be run on the second processor 5003;
  • the second processor 5003 is configured to, when running the computer program, execute:
  • the attribute reconstruction value of the voxel node of the current unit is determined according to the first quantity and the second quantity.
  • the second processor 5003 is further configured to execute any one of the methods described in the foregoing embodiments when running the computer program.
  • the present embodiment provides a decoder in which, when reconstructing the attributes of each voxel node, the judgment conditions for whether to perform attribute encoding and decoding on each voxel node are optimized. Specifically, if there are no duplicate nodes in the current unit, that is, the first number is the same as the second number, then there is no need to encode and decode the voxel nodes of the current unit. Therefore, while ensuring the encoding and decoding efficiency of the point cloud attributes, the time complexity of encoding and decoding the point cloud attributes can be reduced, and the bit rate can also be saved, thereby improving the encoding and decoding performance of the point cloud.
  • a schematic diagram of the composition structure of a coding and decoding system provided in an embodiment of the present application is shown.
  • a coding and decoding system 510 may include an encoder 5101 and a decoder 5102 .
  • the encoder 5101 may be the encoder described in any one of the aforementioned embodiments
  • the decoder 5102 may be the decoder described in any one of the aforementioned embodiments.
  • the first number of voxel nodes of the current unit and the second number of reconstruction nodes of the current unit are determined, and then the attribute reconstruction value of the voxel node of the current unit is determined according to the first number and the second number; wherein the first number and the second number are used to determine whether the voxel node of the current unit skips encoding and decoding.
  • the fourth number of nodes of the current layer and the fifth number of child nodes corresponding to the nodes of the current layer determine the attribute reconstruction value of the child node corresponding to the node of the current layer; wherein the fourth number and the fifth number are used to determine whether the current layer skips encoding and decoding.
  • the number of nodes and the number of child nodes of the current layer can be used to adaptively determine whether the current layer can skip decoding.
  • the time complexity of point cloud attribute encoding and decoding can be reduced, and the bit rate can also be saved, thereby improving the encoding and decoding performance of point cloud.

Landscapes

  • Engineering & Computer Science (AREA)
  • Multimedia (AREA)
  • Signal Processing (AREA)
  • Computing Systems (AREA)
  • Theoretical Computer Science (AREA)
  • Compression Or Coding Systems Of Tv Signals (AREA)

Abstract

本申请实施例公开了一种编解码方法、码流、编码器、解码器以及存储介质，该方法包括：确定当前单元的体素节点的第一数量和当前单元的重建节点的第二数量；其中，第一数量与第二数量用于确定是否对当前单元的体素节点进行跳过解码；根据第一数量与第二数量，确定当前单元的体素节点的属性重建值。这样，在当前单元的体素节点跳过解码时，可以降低属性编解码的时间复杂度，提高点云的属性编解码效率，进而提升点云的编解码性能。

Description

编解码方法、码流、编码器、解码器以及存储介质 技术领域
本申请实施例涉及点云编解码技术领域，尤其涉及一种编解码方法、码流、编码器、解码器以及存储介质。
背景技术
在基于几何的点云压缩(Geometry-based Point Cloud Compression，G-PCC)编解码框架中，点云的几何信息和属性信息是分开进行编码的。其中，G-PCC的属性编码可以包括：预测变换(Predicting Transform，PT)、提升变换(Lifting Transform，LT)以及区域自适应分层变换(Region Adaptive Hierarchical Transform，RAHT)。
在进行RAHT变换中，需要对每一层的节点依次进行变换、预测以及编解码等过程，这样会导致RAHT属性变换编解码的复杂度增大，进而无法有效地去除属性的冗余性，导致属性编码效率较低。
发明内容
本申请实施例提供一种编解码方法、码流、编码器、解码器以及存储介质，能够降低属性编解码的时间复杂度，提高点云的属性编解码效率，进而提升点云的编解码性能。
本申请实施例的技术方案可以如下实现：
第一方面，本申请实施例提供了一种解码方法，应用于解码器，该方法包括：
确定当前单元的体素节点的第一数量和当前单元的重建节点的第二数量；其中，第一数量与第二数量用于确定是否对当前单元的体素节点进行跳过解码；
根据第一数量与第二数量，确定当前单元的体素节点的属性重建值。
第二方面，本申请实施例提供了一种编码方法，应用于编码器，该方法包括：
确定当前单元的体素节点的第一数量和当前单元的重建节点的第二数量；
根据第一数量与第二数量，确定当前单元的体素节点的属性信息是否跳过编码。
第三方面，本申请实施例提供了一种码流，该码流是根据待编码信息进行比特编码生成的；其中，待编码信息包括下述至少一项：
预测模式标识信息、当前单元的重复节点的属性信息和当前层的节点对应的量化后的系数残差；其中，所述预测模式标识信息用于指示所述当前单元是否启动跳过编码模式。
第四方面，本申请实施例提供了一种编码器，该编码器包括第一确定单元和编码单元，其中：
第一确定单元，配置为确定当前单元的体素节点的第一数量和当前单元的重建节点的第二数量；
编码单元，配置为根据第一数量与第二数量，确定当前单元的体素节点的属性信息是否跳过编码。
第五方面，本申请实施例提供了一种编码器，该编码器包括第一存储器和第一处理器，其中：
第一存储器，用于存储能够在第一处理器上运行的计算机程序；
第一处理器，用于在运行计算机程序时，执行如第二方面所述的方法。
第六方面，本申请实施例提供了一种解码器，该解码器包括第二确定单元和第二重建单元，其中：
第二确定单元，配置为确定当前单元的体素节点的第一数量和当前单元的重建节点的第二数量；其中，第一数量与第二数量用于确定是否对当前单元的体素节点进行跳过解码；
第二重建单元，配置为根据第一数量与第二数量，确定当前单元的体素节点的属性重建值。
第七方面，本申请实施例提供了一种解码器，该解码器包括第二存储器和第二处理器，其中：
第二存储器，用于存储能够在第二处理器上运行的计算机程序；
第二处理器，用于在运行计算机程序时，执行如第一方面所述的方法。
第八方面，本申请实施例提供了一种计算机可读存储介质，该计算机可读存储介质存储有计算机程序，所述计算机程序被执行时实现如第一方面所述的方法、或者实现如第二方面所述的方法。
本申请实施例提供了一种编解码方法、码流、编码器、解码器以及存储介质，无论是编码端还是解码端，首先确定当前单元的体素节点的第一数量和当前单元的重建节点的第二数量，然后根据第一数量与第二数量，确定当前单元的体素节点的属性重建值。其中，第一数量与第二数量用于确定当前单元的 体素节点是否跳过编解码。也就是说，在对每个体素节点进行属性重建时，优化了每个体素节点是否进行属性编解码的判断条件，具体为如果当前单元不存在重复节点，即第一数量与第二数量相同，此时不需要编解码当前单元的体素节点，从而在保证点云属性的编解码效率基础上，可以降低点云属性编解码的时间复杂度，并且还可以节省码率，进而提升了点云的编解码性能。
附图说明
图1A为一种三维点云图像示意图；
图1B为一种三维点云图像的局部放大图；
图2A为一种点云图像的六个观看角度示意图；
图2B为一种点云图像对应的数据存储格式示意图；
图3为一种点云编解码的网络架构示意图；
图4A为一种G-PCC编码器的组成框架示意图；
图4B为一种G-PCC解码器的组成框架示意图；
图5A为一种Z轴方向的低平面位置示意图；
图5B为一种Z轴方向的高平面位置示意图；
图6为一种节点编码顺序示意图；
图7A为一种平面标识信息示意图；
图7B为另一种平面标识信息示意图；
图8为一种当前节点的兄弟姐妹节点示意图；
图9为一种激光雷达与节点的相交示意图；
图10为一种处于相同划分深度以及相同坐标的邻域节点示意图；
图11为一种当前节点位于父节点的低平面位置示意图；
图12为一种当前节点位于父节点的高平面位置示意图；
图13为一种激光雷达点云平面位置信息的预测编码示意图；
图14为一种IDCM编码示意图；
图15为一种旋转激光雷达获取点云的坐标转换示意图；
图16为一种X轴或Y轴方向的预测编码示意图；
图17A为一种通过水平方位角来进行预测Y平面的角度示意图；
图17B为一种通过水平方位角来进行预测X平面的角度示意图；
图18为另一种X轴或Y轴方向的预测编码示意图；
图19A为一种子块包括的三个交点示意图；
图19B为一种利用三个交点拟合的三角面片集示意图；
图19C为一种三角面片集的上采样示意图；
图20为一种基于距离的LOD构造过程的示意图；
图21为一种LOD生成过程的可视化结果示意图；
图22为一种属性预测的编码流程示意图；
图23为一种金字塔结构的组成示意图；
图24为另一种金字塔结构的组成示意图；
图25为一种层间最近邻查找的LOD结构示意图；
图26为一种基于空间关系进行最近邻查找结构示意图；
图27A为一种共面的空间关系示意图；
图27B为一种共面和共线的空间关系示意图；
图27C为一种共面、共线和共点的空间关系示意图；
图28为一种基于快速查找的层间预测示意图；
图29为一种属性层内最近邻查找的LOD结构示意图；
图30为一种基于快速查找的层内预测示意图；
图31为一种基于块进行邻域查找结构示意图；
图32为一种提升变换的编码流程示意图；
图33为一种RAHT变换结构示意图；
图34为一种RAHT沿x、y、z三方向的变换过程示意图；
图35A为一种RAHT正变换的过程示意图；
图35B为一种RAHT逆变换的过程示意图；
图36为一种属性编码块的结构示意图；
图37为一种RAHT属性预测变换编码的整体流程示意图；
图38为一种当前块的邻域预测关系示意图；
图39为一种属性变换系数的计算过程示意图；
图40为一种RAHT属性帧间预测编码的结构示意图；
图41为本申请实施例提供的一种解码方法的流程示意图一；
图42为本申请实施例提供的一种RAHT属性编码层的结构示意图；
图43为本申请实施例提供的一种解码方法的流程示意图二；
图44为本申请实施例提供的一种编码方法的流程示意图一；
图45为本申请实施例提供的一种编码方法的流程示意图二；
图46为本申请实施例提供的一种编码方法的流程示意图三；
图47为本申请实施例提供的一种编码器的组成结构示意图；
图48为本申请实施例提供的一种编码器的具体硬件结构示意图；
图49为本申请实施例提供的一种解码器的组成结构示意图；
图50为本申请实施例提供的一种解码器的具体硬件结构示意图；
图51为本申请实施例提供的一种编解码系统的组成结构示意图。
具体实施方式
为了能够更加详尽地了解本申请实施例的特点与技术内容，下面结合附图对本申请实施例的实现进行详细阐述，所附附图仅供参考说明之用，并非用来限定本申请实施例。
除非另有定义，本文所使用的所有的技术和科学术语与属于本申请的技术领域的技术人员通常理解的含义相同。本文中所使用的术语只是为了描述本申请实施例的目的，不是旨在限制本申请。
在以下的描述中，涉及到“一些实施例”，其描述了所有可能实施例的子集，但是可以理解，“一些实施例”可以是所有可能实施例的相同子集或不同子集，并且可以在不冲突的情况下相互结合。
还需要指出，本申请实施例所涉及的术语“第一\第二\第三”仅是用于区别类似的对象，不代表针对对象的特定排序，可以理解地，“第一\第二\第三”在允许的情况下可以互换特定的顺序或先后次序，以使这里描述的本申请实施例能够以除了在这里图示或描述的以外的顺序实施。
点云(Point Cloud)是物体表面的三维表现形式，通过光电雷达、激光雷达、激光扫描仪、多视角相机等采集设备，可以采集得到物体表面的点云(数据)。
点云是空间中一组无规则分布的、表达三维物体或场景的空间结构及表面属性的离散点集，图1A展示了三维点云图像和图1B展示了三维点云图像的局部放大图，可以看到点云表面是由分布稠密的点所组成的。
二维图像在每一个像素点均有信息表达，分布规则，因此不需要额外记录其位置信息；然而点云中的点在三维空间中的分布具有随机性和不规则性，因此需要记录每一个点在空间中的位置，才能完整地表达一幅点云。与二维图像类似，采集过程中每一个位置均有对应的属性信息，通常为RGB颜色值，颜色值反映物体的色彩；对于点云来说，每一个点所对应的属性信息除了颜色信息以外，还有比较常见的是反射率(reflectance)值，反射率值反映物体的表面材质。因此，点云数据通常包括点的位置信息和点的属性信息。其中，点的位置信息也可称为点的几何信息。例如，点的几何信息可以是点的三维坐标信息(x,y,z)。点的属性信息可以包括颜色信息和/或反射率等等。例如，反射率可以是一维反射率信息(r)；颜色信息可以是任意一种色彩空间上的信息，或者颜色信息也可以是三维颜色信息，如RGB信息。在这里，R表示红色(Red，R)，G表示绿色(Green，G)，B表示蓝色(Blue，B)。再如，颜色信息可以是亮度色度(YCbCr，YUV)信息。其中，Y表示明亮度(Luma)，Cb(U)表示蓝色色差，Cr(V)表示红色色差。
根据激光测量原理得到的点云，点云中的点可以包括点的三维坐标信息和点的反射率值。再如，根据摄影测量原理得到的点云，点云中的点可以可包括点的三维坐标信息和点的三维颜色信息。再如，结合激光测量和摄影测量原理得到点云，点云中的点可以可包括点的三维坐标信息、点的反射率值和点的三维颜色信息。
如图2A和图2B所示为一幅点云图像及其对应的数据存储格式。其中，图2A提供了点云图像的六个观看角度，图2B由文件头信息部分和数据部分组成，头信息包含了数据格式、数据表示类型、点云总点数、以及点云所表示的内容。例如，点云为“.ply”格式，由ASCII码表示，总点数为207242，每 个点具有三维坐标信息(x,y,z)和三维颜色信息(r,g,b)。
点云可以按获取的途径分为：
静态点云：即物体是静止的，获取点云的设备也是静止的；
动态点云：物体是运动的，但获取点云的设备是静止的；
动态获取点云：获取点云的设备是运动的。
例如，按点云的用途分为两大类：
类别一：机器感知点云，其可以用于自主导航系统、实时巡检系统、地理信息系统、视觉分拣机器人、抢险救灾机器人等场景；
类别二：人眼感知点云，其可以用于数字文化遗产、自由视点广播、三维沉浸通信、三维沉浸交互等点云应用场景。
点云可以灵活方便地表达三维物体或场景的空间结构及表面属性，并且由于点云通过直接对真实物体采样获得，在保证精度的前提下能提供极强的真实感，因而应用广泛，其范围包括虚拟现实游戏、计算机辅助设计、地理信息系统、自动导航系统、数字文化遗产、自由视点广播、三维沉浸远程呈现、生物组织器官三维重建等。
点云的采集主要有以下途径：计算机生成、3D激光扫描、3D摄影测量等。计算机可以生成虚拟三维物体及场景的点云；3D激光扫描可以获得静态现实世界三维物体或场景的点云，每秒可以获取百万级点云；3D摄影测量可以获得动态现实世界三维物体或场景的点云，每秒可以获取千万级点云。这些技术降低了点云数据获取成本和时间周期，提高了数据的精度。点云数据获取方式的变革，使大量点云数据的获取成为可能，伴随着应用需求的增长，海量3D点云数据的处理遭遇存储空间和传输带宽限制的瓶颈。
示例性地，以帧率为30帧每秒(fps)的点云视频为例，每帧点云的点数为70万，每个点具有坐标信息xyz(float)和颜色信息RGB(uchar)，则10s点云视频的数据量大约为0.7million×(4Byte×3+1Byte×3)×30fps×10s＝3.15GB，其中，1Byte为10bit；而YUV采样格式为4:2:0，帧率为24fps的1280×720二维视频，其10s的数据量约为1280×720×12bit×24fps×10s≈0.33GB，10s的两视角三维视频的数据量约为0.33×2＝0.66GB。由此可见，点云视频的数据量远超过相同时长的二维视频和三维视频的数据量。因此，为更好地实现数据管理，节省服务器存储空间，降低服务器与客户端之间的传输流量及传输时间，点云压缩成为促进点云产业发展的关键问题。
也就是说，由于点云是海量点的集合，存储点云不仅会消耗大量的内存，而且不利于传输，也没有这么大的带宽可以支持将点云不经过压缩直接在网络层进行传输，因此，需要对点云进行压缩。
目前，可对点云进行压缩的点云编码框架可以是运动图像专家组(Moving Picture Experts Group，MPEG)提供的基于几何的点云压缩(Geometry-based Point Cloud Compression，G-PCC)编解码框架或基于视频的点云压缩(Video-based Point Cloud Compression，V-PCC)编解码框架，也可以是AVS提供的AVS-PCC编解码框架。G-PCC编解码框架可用于针对第一类静态点云和第三类动态获取点云进行压缩，其可以是基于点云压缩测试平台(Test Model Compression 13，TMC13)，V-PCC编解码框架可用于针对第二类动态点云进行压缩，其可以是基于点云压缩测试平台(Test Model Compression 2，TMC2)。故G-PCC编解码框架也称为点云编解码器TMC13，V-PCC编解码框架也称为点云编解码器TMC2。
本申请实施例提供了一种包含解码方法和编码方法的点云编解码系统的网络架构，图3为本申请实施例提供的一种点云编解码的网络架构示意图。如图3所示，该网络架构包括一个或多个电子设备13至1N和通信网络01，其中，电子设备13至1N可以通过通信网络01进行视频交互。电子设备在实施的过程中可以为各种类型的具有点云编解码功能的设备，例如，所述电子设备可以包括手机、平板电脑、个人计算机、个人数字助理、导航仪、数字电话、视频电话、电视机、传感设备、服务器等，本申请实施例不作限制。其中，本申请实施例中的解码器或编码器就可以为上述电子设备。
其中，本申请实施例中的电子设备具有点云编解码功能，一般包括点云编码器(即编码器)和点云解码器(即解码器)。
下面以G-PCC编解码框架为例进行相关技术的说明。
可以理解，在点云G-PCC编解码框架中，针对待编码的点云数据，首先通过片(slice)划分，将点云数据划分为多个slice。在每一个slice中，点云的几何信息和每个点所对应的属性信息是分开进行编码的。
图4A示出了一种G-PCC编码器的组成框架示意图。如图4A所示，在几何编码过程中，对几何信息进行坐标转换，使点云全都包含在一个包围盒(Bounding Box)中，然后再进行量化，这一步量化主要起到缩放的作用，由于量化取整，使得一部分点云的几何信息相同，于是再基于参数来决定是否移除重复点，量化和移除重复点这一过程又被称为体素化过程。接着对包围盒进行八叉树划分或者预测树构 建。在该过程中，针对划分的叶子结点中的点进行算术编码，生成二进制的几何比特流；或者，针对划分产生的交点(Vertex)进行算术编码(基于交点进行表面拟合)，生成二进制的几何比特流。在属性编码过程中，几何编码完成，对几何信息进行重建后，需要先进行颜色转换，将颜色信息(即属性信息)从RGB颜色空间转换到YUV颜色空间。然后，利用重建的几何信息对点云重新着色，使得未编码的属性信息与重建的几何信息对应起来。属性编码主要针对颜色信息进行，在颜色信息编码过程中，主要有两种变换方法，一是依赖于细节层次(Level of Detail，LOD)划分的基于距离的提升变换，二是直接进行区域自适应分层变换(Region Adaptive Hierarchal Transform，RAHT)，这两种方法都会将颜色信息从空间域转换到频域，通过变换得到高频系数和低频系数，最后对系数进行量化，再对量化系数进行算术编码，可以生成二进制的属性比特流。
图4B示出了一种G-PCC解码器的组成框架示意图。如图4B所示，针对所获取的二进制比特流，首先对二进制比特流中的几何比特流和属性比特流分别进行独立解码。在对几何比特流的解码时，通过算术解码-重构八叉树/重构预测树-重建几何-坐标逆转换，得到点云的几何信息；在对属性比特流的解码时，通过算术解码-反量化-LOD划分/RAHT-颜色逆转换，得到点云的属性信息，基于几何信息和属性信息还原待编码的点云数据(即输出点云)。
需要说明的是，在如图4A或图4B所示，目前G-PCC的几何编解码可以分为基于八叉树的几何编解码(用虚线框标识)和基于预测树的几何编解码(用点划线框标识)。
对于基于八叉树的几何编码(Octree geometry encoding，OctGeomEnc)而言，基于八叉树的几何编码包括：首先对几何信息进行坐标转换，使点云全都包含在一个包围盒中。然后再进行量化，这一步量化主要起到缩放的作用，由于量化取整，使得一部分点的几何信息相同，根据参数来决定是否移除重复点，量化和移除重复点这一过程又被称为体素化过程。接下来，按照广度优先遍历的顺序不断对包围盒进行树划分(例如八叉树、四叉树、二叉树等)，对每个节点的占位码进行编码。在相关技术中，某公司提出了一种隐式几何的划分方式，首先计算点云的包围盒假设dx>dy>dz，该包围盒对应为一个长方体。在几何划分时，首先会基于x轴一直进行二叉树划分，得到两个子节点；直到满足dx＝dy>dz条件时，才会基于x和y轴一直进行四叉树划分，得到四个子节点；当最终满足dx＝dy＝dz条件时，会一直进行八叉树划分，直到划分得到的叶子结点为1×1×1的单位立方体时停止划分，对叶子结点中的点进行编码，生成二进制码流。在基于二叉树/四叉树/八叉树划分的过程中，引入两个参数：K、M。参数K指示在进行八叉树划分之前二叉树/四叉树划分的最多次数；参数M用来指示在进行二叉树/四叉树划分时对应的最小块边长为2M。同时K和M必须满足条件：假设dmax＝max(dx,dy,dz)，dmin＝min(dx,dy,dz)，参数K满足：K≥dmax-dmin；参数M满足：M≥dmin。参数K与M之所以满足上述的条件，是因为目前G-PCC在几何隐式划分的过程中，划分方式的优先级为二叉树、四叉树和八叉树，当节点块大小不满足二叉树/四叉树的条件时，才会对节点一直进行八叉树的划分，直到划分到叶子节点最小单位1×1×1。基于八叉树的几何信息编码模式可以通过利用空间中相邻点之间的相关性来对点云的几何信息进行有效的编码，但是对于一些较为平坦的节点或者具有平面特性的节点，通过利用平面编码可以进一步提升点云几何信息的编码效率。
示例性地，图5A和图5B提供了一种平面位置示意图。其中，图5A示出了一种Z轴方向的低平面位置示意图，图5B示出了一种Z轴方向的高平面位置示意图。如图5A所示，这里的(a)、(a0)、(a1)、(a2)、(a3)均属于Z轴方向的低平面位置，以(a)为例，可以看到当前节点中被占据的四个子节点都位于当前节点在Z轴方向的低平面位置，那么可以认为当前节点属于一个Z平面并且在Z轴方向是一个低平面。同理，如图5B所示，这里的(b)、(b0)、(b1)、(b2)、(b3)均属于Z轴方向的高平面位置，以(b)为例，可以看到当前节点中被占据的四个子节点位于当前节点在Z轴方向的高平面位置，那么可以认为当前节点属于一个Z平面并且在Z轴方向是一个高平面。
进一步地，以图5A中的(a)为例，对八叉树编码和平面编码效率进行比较，图6提供了一种节点编码顺序示意图，即按照图6所示的0、1、2、3、4、5、6、7的顺序进行节点编码。在这里，如果对图5A中的(a)采用八叉树编码方式，那么当前节点的占位信息表示为：11001100。但是如果采用平面编码方式，首先需要编码一个标识符表示当前节点在Z轴方向是一个平面，其次如果当前节点在Z轴方向是一个平面，还需要对当前节点的平面位置进行表示；其次仅仅需要对Z轴方向的低平面节点的占位信息进行编码(即0、2、4、6四个子节点的占位信息)，因此基于平面编码方式对当前节点进行编码，仅仅需要编码6个比特(bit)，相比相关技术的八叉树编码可以减少2个bit的表示。基于此分析，平面编码相比八叉树编码具有较为明显的编码效率。因此，对于一个被占据的节点，如果在某一个维度上采用平面编码方式进行编码，首先需要对当前节点在该维度上的平面标识(planarMode)和平面位置(PlanePos)信息进行表示，其次基于当前节点的平面信息来对当前节点的占位信息进行编码。示例性地，图7A示出了一种平面标识信息示意图。如图7A所示，这里在Z轴方向为一个低平面；对应 地，平面标识信息的取值为真(true)或者1，即planarMode_Z＝true；平面位置信息为低平面(low)，即PlanePosition_Z＝low。图7B示出了另一种平面标识信息示意图。如图7B所示，这里在Z轴方向不为一个平面；对应地，平面标识信息的取值为假(false)或者0，即planarMode_Z＝false。
需要注意的是，对于PlaneMode_i：0代表当前节点在i轴方向不是一个平面，1代表当前节点在i轴方向是一个平面。若当前节点在i轴方向是一个平面，则对于PlanePosition_i：0代表当前节点在i轴方向是一个平面，并且平面位置为低平面，1表示当前节点在i轴方向上是一个高平面。其中，i表示坐标维度，可以为X轴方向、Y轴方向或者Z轴方向，故i＝0,1,2。
在G-PCC标准中，判断一个节点是否满足平面编码的条件以及在该节点满足平面编码条件时，需要对该节点的平面标识和平面位置信息的预测编码。
在本申请实施例中，当前G-PCC标准中存在三种判断节点是否满足平面编码的判断条件，下面对其逐一进行详细说明。
一、根据节点在每个维度上的平面概率进行判断。
(1)确定当前节点的局部区域密度(local_node_density)；
(2)确定当前节点在每个维度上的概率Prob(i)。
在节点的局部区域密度小于阈值Th(例如Th＝3)时，利用当前节点在三个坐标维度上的平面概率Prob(i)和阈值Th0、Th1和Th2进行比较，其中Th0<Th1<Th2(例如，Th0＝0.6，Th1＝0.77，Th2＝0.88)，这里可以利用Eligiblei(i＝0,1,2)表示每个维度上是否启动平面编码：Eligiblei＝Prob(i)>＝threshold。
需要注意的是，threshold是进行自适应变化的，例如，当Prob(0)>Prob(1)>Prob(2)时，则Eligiblei的设置如下：
Eligible0＝Prob(0)>＝Th0；
Eligible1＝Prob(1)>＝Th1；
Eligible2＝Prob(2)>＝Th2。
当Prob(1)>Prob(0)>Prob(2)时，则Eligiblei的设置如下：
Eligible0＝Prob(0)>＝Th1；
Eligible1＝Prob(1)>＝Th0；
Eligible2＝Prob(2)>＝Th2。
在这里，Prob(i)的更新具体如下：
Prob(i)new＝(L×Prob(i)+δ(coded node))/L+1  (1)
其中，L＝255；另外，若coded node节点是一个平面，则δ(coded node)为1；否则δ(coded node)为0。
在这里，local_node_density的更新具体如下：
local_node_densitynew＝local_node_density+4*numSiblings  (2)
其中，local_node_density初始化为4，numSiblings为该节点的兄弟姐妹节点数目。示例性地，图8示出了一种当前节点的兄弟姐妹节点示意图。如图8所示，当前节点为用斜线填充的节点，用网格填充的节点为兄弟姐妹节点，那么当前节点的兄弟姐妹节点数目为5(包括当前节点自身)。
二、根据当前层的点云密度来判断当前层节点是否满足平面编码。
利用当前层点的密度来判断是否对当前层的节点进行平面编码。假设当前待编码点云的点数为pointCount，经过推断直接编码模式(Infer Direct Coding Model，IDCM)编码已经重建出的点数为numPointCountRecon，又因为八叉树是基于广度优先遍历的顺序进行编码，因此可以得到当前层待编码的节点数目假设为nodeCount，那么判断当前层是否启动平面编码假设为planarEligibleKOctreeDepth，具体为：planarEligibleK OctreeDepth＝(pointCount-numPointCountRecon)<nodeCount×1.3。
其中，若(pointCount-numPointCountRecon)小于nodeCount×1.3，则planarEligibleK OctreeDepth为true；若(pointCount-numPointCountRecon)不小于nodeCount×1.3，则planarEligibleKOctreeDepth为false。这样，当planarEligibleKOctreeDepth为true时，则在当前层所有节点都进行平面编码；否则在当前层所有节点都不进行平面编码，仅仅采用八叉树编码。
三、根据激光雷达点云的采集参数来判断当前节点是否满足平面编码。
图9示出了一种激光雷达与节点的相交示意图。如图9所示，用网格填充的节点同时被两个激光射线(Laser)穿过，因此当前节点在Z轴垂直方向上不是一个平面；用斜线填充的节点足够小到不能同时被两个Laser同时穿过，因此斜线填充的节点在Z轴垂直方向上有可能是一个平面。
进一步地，针对满足平面编码条件的节点，可以对平面标识信息和平面位置信息进行预测编码。
首先，平面标识信息的预测编码。
在这里，仅仅采用三个上下文信息进行编码，即各个坐标维度上的平面标识分开进行上下文设计。
其次，平面位置信息的预测编码。
应理解，针对非激光雷达点云平面位置信息的编码而言，平面位置信息的预测编码可以包括：
(a)利用邻域节点的占位信息进行预测得到当前节点的平面位置信息为三元素：预测为低平面、预测为高平面和无法预测；
(b)与当前节点在相同划分深度以及相同坐标下的节点与当前节点之间的空间距离：“近”和“远”；
(c)与当前节点在相同划分深度以及相同坐标下的节点如果是一个平面，则确定该节点的平面位置；
(d)坐标维度(i＝0,1,2)。
需要说明的是，在本申请实施例中，确定出与当前节点在相同划分深度以及相同坐标下的节点和当前节点之间的空间距离之后，如果该空间距离小于预设距离阈值，那么可以确定该空间距离为“近”；或者，如果该空间距离大于预设距离阈值，那么可以确定该空间距离为“远”。
示例性地，图10示出了一种处于相同划分深度以及相同坐标的邻域节点示意图。如图10所示，加粗的大立方体表示父节点(Parent node)，其内部网格填充的小立方体表示当前节点(Current node)，并且示出了当前节点的交点位置(Vertex position)；白色填充的小立方体表示处于相同划分深度以及相同坐标的邻域节点，当前节点与邻域节点之间的距离为空间距离，可以判断为“近”或“远”；另外，如果该邻域节点为一个平面，那么还需要该邻域节点的平面位置(Planar position)。
这样，如图10所示，当前节点为网格填充的小立方体，则在相同的八叉树划分深度等级下，以及相同的垂直坐标下查找邻域节点为白色填充的小立方体，判断两个节点之间的距离为“近”和“远”，并且参考节点的平面位置。
进一步地，在本申请实施例中，图11示出了一种当前节点位于父节点的低平面位置示意图。如图11所示，(a)、(b)、(c)示出了三种当前节点位于父节点的低平面位置的示例。具体说明如下：
①如果点填充节点的子节点4到7中有任何一个被占用，而所有网格填充节点都未被占用，则极有可能在当前节点(用斜线填充)中存在一个平面，且该平面位置较低。
②如果点填充节点的子节点4到7都未被占用，而任何网格填充节点被占用，则极有可能在当前节点(用斜线填充)中存在一个平面，且该平面位置较高。
③如果点填充节点的子节点4到7均为空节点，网格填充节点均为空节点，则无法推断平面位置，故标记为未知。
④如果点填充节点的子节点4到7中有任何一个被占用，而网格填充节点中有任何一个被占用，此时也无法推断出平面位置，因此将其标记为未知。
在本申请实施例中，图12示出了一种当前节点位于父节点的高平面位置示意图。如图12所示，(a)、(b)、(c)示出了三种当前节点位于父节点的高平面位置的示例。具体说明如下：
①如果网格填充节点的子节点4到7中有任何一个节点被占用，而点填充节点未被占用，则极有可能在当前节点(用斜线填充)中存在一个平面，且平面位置较低。
②如果网格填充节点的子节点4到7均未被占用，而点填充节点被占用，则极有可能在当前节点(用斜线填充)中存在平面，且平面位置较高。
③如果网格填充节点的子节点4到7都是未被占用的，而点填充节点是未被占用的，此时无法推断平面位置，因此标记为未知。
④如果网格填充节点的子节点4到7中有一个被占用，而点填充节点被占用，此时无法推断平面位置，因此标记为未知。
还应理解，针对激光雷达点云平面位置信息的编码而言，图13示出了一种激光雷达点云平面位置信息的预测编码示意图。如图13所示，在激光雷达的发射角度为θbottom时，这时候可以映射为低平面(Bottom virtual plane)；在激光雷达的发射角度为θtop时，这时候可以映射为高平面(Top virtual plane)。
也就是说，通过利用激光雷达采集参数来预测当前节点的平面位置，通过利用当前节点与激光射线相交的位置来将位置量化为多个区间，最终作为当前节点平面位置的上下文信息。具体计算过程如下：假设激光雷达的坐标为(xLidar,yLidar,zLidar)，当前节点的几何坐标为(x,y,z)，那么首先计算当前节点相对于激光雷达的垂直正切值tanθ，计算公式如下：
进一步地，又因为每个Laser会相对于激光雷达有一定偏移角度，因此还需要计算当前节点相对于Laser的相对正切值tanθcorr,L，具体计算如下：
最终会利用当前节点的相对正切值tanθcorr,L来对当前节点的平面位置进行预测，具体如下，假设当 前节点下边界的正切值为tan(θbottom)，上边界的正切值为tan(θtop)，根据tanθcorr,L将平面位置量化为4个量化区间，即确定平面位置的上下文信息。
但是基于八叉树的几何信息编码模式仅对空间中具有相关性的点有高效的压缩速率，而对于在几何空间中处于孤立位置的点来说，使用直接编码模式(Direct Coding Model，DCM)可以大大降低复杂度。对于八叉树中的所有节点，DCM的使用不是通过标志位信息来表示的，而是通过当前节点父节点和邻居信息来进行推断得到。判断当前节点是否具有DCM编码资格的方式有三种，具体如下：
(1)当前节点没有兄弟姐妹子节点，即当前节点的父节点只有一个孩子节点，同时当前节点父节点的父节点仅有两个被占据子节点，即当前节点最多只有一个邻居节点。
(2)当前节点的父节点仅有当前节点一个占据子节点，同时与当前节点共用一个面的六个邻居节点也都属于空节点。
(3)当前节点的兄弟姐妹节点数目大于1。
示例性地，图14提供了一种IDCM编码示意图。如果当前节点不具有DCM编码资格将对其进行八叉树划分，若具有DCM编码资格将进一步判断该节点中包含的点数，当点数小于阈值(例如2)时，则对该节点进行DCM编码，否则将继续进行八叉树划分。当应用DCM编码模式时，首先需要编码当前节点是否是一个真正的孤立点，即IDCM_flag，当IDCM_flag为true时，则当前节点采用DCM编码，否则仍然采用八叉树编码。在当前节点满足DCM编码时，需要编码当前节点的DCM编码模式，目前存在两种DCM模式，分别是：(a)仅仅只有一个点存在(或者是多个点，但是属于重复点)；(b)含有两个点。最后需要编码每个点的几何信息，假设节点的边长为2d时，对该节点几何坐标的每一个分量进行编码时需要d比特，该比特信息直接被编进码流中。这里需要注意的是，在对激光雷达点云进行编码时，通过利用激光雷达采集参数来对三个维度的坐标信息进行预测编码，从而可以进一步提升几何信息的编码效率。
进一步地，下面针对IDCM编码的过程进行详细介绍。
当前节点满足DCM编码模式时，首先编码当前节点的点数目numPoints；根据不同的DirectMode来对当前节点的点数目进行编码：
(1)如果当前节点不满足DCM节点的要求，则直接退出(即点数大于2个点，并且不是重复点)。
(2)当前节点含有的点数numPonts小于或等于2，则编码过程如下：
i)首先编码当前节点的numPonts是否大于1；
ii)如果当前节点只有一个点并且几何编码环境为几何无损编码，则需要编码当前节点的第二个点不是重复点。
(3)当前节点含有的点数numPonts大于2，则编码过程如下：
i)首先编码当前节点的numPonts小于或等于1；
ii)其次编码当前节点的第二个点是一个重复点，其次编码当前节点的重复点数目是否大于1，当重复点数目大于1时，需要对剩余的重复点数目进行指数哥伦布解码。
在编码完成当前节点的点数目之后，对当前节点中包含点的坐标信息进行编码。下面将分别对激光雷达点云和面向人眼点云进行详细介绍。
(一)面向人眼点云。
(1)如果当前节点中仅仅只含有一个点，则会对点的三个维度方向的几何信息进行直接编码(Bypass coding)；
(2)如果当前节点中含有两个点，则会首先通过利用点的几何坐标得到优先编码的坐标轴dirextAxis。这里需要注意的是，目前比较的坐标轴只包含x轴和y轴，不包含z轴。假设当前节点的几何坐标为nodePos，则判断的方式如下：
dirextAxis＝！(nodePos[0]<nodePos[1])  (5)
也就是会将节点坐标几何位置小的轴作为优先编码的坐标轴dirextAxis，其次按照如下方式首先对优先编码的坐标轴dirextAxis几何信息进行编码。假设优先编码的轴对应的代编码几何bit深度为nodeSizeLog2，并假设两个点的坐标分别为pointPos[0]和pointPos[1]。具体编码过程如下：

在编码完成优先编码的坐标轴dirextAxis之后，再继续对当前节点的几何坐标进行直接编码。假设每个点的剩余编码bit深度为nodeSizeLog2，则具体编码过程如下：
for(int axisIdx＝0；axisIdx<3；++axisIdx)
for(int mask＝(1<<nodeSizeLog2[axisIdx])>>1；mask；mask>>1)
encodePosBit(！！(pointPos[axisIdx]&mask))。
(二)面向激光雷达点云。
如果当前节点中含有两个点，则会首先通过利用点的几何坐标得到优先编码的坐标轴dirextAxis，假设当前节点的几何坐标为nodePos，则判断的方式如下：
dirextAxis＝！(nodePos[0]<nodePos[1])
也就是会将节点坐标几何位置小的轴作为优先编码的坐标轴dirextAxis，这里需要注意的是，目前比较的坐标轴只包含x轴和y轴，不包含z轴。其次按照如下方式首先对优先编码的坐标轴dirextAxis几何信息进行编码，假设优先编码的轴对应的代编码几何bit深度为nodeSizeLog2，并假设两个点的坐标分别为pointPos[0]和pointPos[1]。具体编码过程如下：
在编码完成优先编码的坐标轴dirextAxis之后，再对当前节点的几何坐标进行编码。
由于激光雷达点云可以得到激光雷达点云的采集参数，通过利用可以预测当前节点的几何坐标信息，从而可以进一步提升点云的几何信息编码效率。同样的首先利用当前节点的几何信息nodePos得到一个直接编码的主轴方向，其次利用已经完成编码的方向的几何信息来对另外一个维度的几何信息进行预测编码。同样假设直接编码的轴方向是directAxis，并且假设直接编码中的代编码bit深度为nodeSizeLog2，则编码方式如下：
for(int mask＝(1<<nodeSizeLog2)>>1；mask；mask>>1)；
encodePosBit(！！(pointPos[directAxis]&mask))。
这里需要注意的是，在这里会将directAxis方向的几何精度信息全部编码。
示例性地，图15提供了一种旋转激光雷达获取点云的坐标转换示意图。其中，在笛卡尔坐标系下，对于每一个节点的(x,y,z)坐标，均可以转换为用(R,i)表示。另外，激光扫描器(Laser Scanner)可以按照预设角度进行激光扫描，在i的不同取值下，可以得到不同的θ(i)。例如，在i等于1时，这时候可以得到θ(1)，对应的扫描角度为-15°；在i等于2时，这时候可以得到θ(2)，对应的扫描角度为-13°；在i等于10时，这时候可以得到θ(10)，对应的扫描角度为+13°；在i等于9时，这时候可以得到θ(19)，对应的扫描角度为+15°。
这样，在编码完成directAxis坐标方向的所有精度之后，会首先计算当前点所对应的LaserIdx，即图15中的pointLaserIdx号，并且计算当前节点的LaserIdx，即nodeLaserIdx；其次会利用节点的LaserIdx即nodeLaserIdx来对点的LaserIdx即pointLaserIdx进行预测编码，其中节点或者点的LaserIdx的计算方式如下。假设点的几何坐标为pointPos，激光射线的起始坐标为LidarOrigin，并且假设Laser的数目为LaserNum，每个Laser的正切值为tanθi，每个Laser在垂直方向上的偏移位置为Zi，则：

在计算得到当前点的LaserIdx之后，首先会利用当前节点的LaserIdx对点的pointLaserIdx进行预测编码。在编码完成当前点的LaserIdx之后，对当前点三个维度的几何信息利用激光雷达的采集参数进行预测编码。
示例性地，图16示出了一种X轴或Y轴方向的预测编码示意图。如图16所示，用网格填充的方框表示当前点(Current node)，用斜线填充的方框表示已编码点(Already coded node)。在这里，首先利用当前点对应的LaserIdx得到对应的水平方位角的预测值，即其次利用当前点对应的节点几何信息得到节点对应的水平方位角度其中，假设节点的几何坐标为nodePos，则水平方位角与节点几何信息之间的计算方式如下：
通过利用激光雷达的采集参数，可以得到每个Laser的旋转点数numPoints，即代表每个激光射线旋转一圈得到的点数，则可以利用每个Laser的旋转点数计算得到每个Laser的旋转角速度deltaPhi，计算方式如下：
进一步地，利用节点的水平方位角以及当前点对应的Laser前一个编码点的水平方位角计算得到当前点对应的水平方位角预测值即如图17A和图17B所示的水平方位角的预测值。其中，图17A示出了一种通过水平方位角来进行预测Y平面的角度示意图，图17B示出了一种通过水平方位角来进行预测X平面的角度示意图。在这里，对于当前点对应的水平方位角预测值计，算方式如下：
示例性地，图18示出了另一种X轴或Y轴方向的预测编码示意图。如图18所示，用网格填充的部分(左侧)表示低平面，用点填充的部分(右侧)表示高平面，表示当前节点的低平面水平方位角，表示当前节点的高平面水平方位角，表示当前节点对应的水平方位角预测值。
这样，通过利用水平方位角的预测值以及当前节点的低平面水平方位角和高平面水平方位角来对当前节点的几何信息进行预测编码。具体如下所示：

int context＝(angLel≥0&&angLeR≥0)||(angLel<0&&angLeR<0)？0:2；
int minAngle＝std∷min(abs(angLel),abs(angLeR))；
int maxAngle＝std∷max(abs(angLel),abs(angLeR))；
context+＝maxAngle>minAngle？0:1；
context+＝maxAngle>minAngle？0:4。
在编码完成点的LaserIdx之后，会利用当前点所对应的LaserIdx对当前点的Z轴方向进行预测编码，即当前通过利用当前点的x和y信息计算得到雷达坐标系的深度信息radius，其次利用当前点的激光LaserIdx得到当前点的正切值以及垂直方向的偏移量，则可以得到当前点的Z轴方向的预测值即Z_pred。具体如下所示：
int tanTheta＝tanθlaserIdx；
int zOffset＝ZlaserIdx；
Z_pred＝radius×tanTheta-zOffset。
进一步地，利用Z_pred对当前点的Z轴方向的几何信息进行预测编码得到预测残差Z_res，最终对Z_res进行编码。
需要注意的是，在节点划分到叶子节点时，在几何无损编码的情况下，需要对叶子节点中的重复点数目进行编码。最终对所有节点的占位信息进行编码，生成二进制码流。另外G-PCC目前引入了一种平面编码模式，在对几何进行划分的过程中，会判断当前节点的子节点是否处于同一平面，如果当前节点的子节点满足同一平面的条件，会用该平面对当前节点的子节点进行表示。
对于基于八叉树的几何解码而言，解码端按照广度优先遍历的顺序，在对每个节点的占位信息解码之前，首先会利用已经重建得到的几何信息来判断当前节点是否进行平面解码或者IDCM解码，如果 当前节点满足平面解码的条件，则会首先对当前节点的平面标识和平面位置信息进行解码，其次基于平面信息来对当前节点的占位信息进行解码；如果当前节点满足IDCM解码的条件，则会首先解码当前节点是否是一个真正的IDCM节点，如果是一个真正的IDCM解码，则会继续解析当前节点的DCM解码模式，其次可以得到当前DCM节点中的点数目，最后对每个点的几何信息进行解码。对于既不满足平面解码也不满足DCM解码的节点，会对当前节点的占位信息进行解码。通过按照这样的方式不断解析得到每个节点的占位码，并且依次不断划分节点，直至划分得到1×1×1的单位立方体时停止划分，解析得到每个叶子节点中包含的点数，最终恢复得到几何重构点云信息。
下面对IDCM解码的过程进行详细介绍。
与编码端的处理过程类似，首先利用先验信息来决定节点是否启动IDCM，即IDCM的启动条件如下：
(1)当前节点没有兄弟姐妹子节点，即当前节点的父节点只有一个孩子节点，同时当前节点父节点的父节点仅有两个被占据子节点，即当前节点最多只有一个邻居节点。
(2)当前节点的父节点仅有当前节点一个占据子节点，同时与当前节点共用一个面的六个邻居节点也都属于空节点。
(3)当前节点的兄弟姐妹节点数目大于1。
进一步地，当节点满足DCM编码的条件时，首先解码当前节点是否是一个真正的DCM节点，即IDCM_flag；当IDCM_flag为true时，则当前节点采用DCM编码，否则仍然采用八叉树编码。
其次解码当前节点的点数目numPoints，具体的解码方式如下所示：
i)首先解码当前节点的numPonts是否大于1；
ii)如果解码得到当前节点的numPonts大于1，则继续解码第二个点是否是一个重复点；如果第二个点不是重复点，则这里可以隐性推断出满足DCM模式的第二种，只含有两个点；
iii)如果解码得到当前节点的numPonts小于等于1，则继续解码第二个点是否是一个重复点；如果第二个点不是重复点，则这里可以隐性推断出满足DCM模式的第二种，只含有一个点；如果解码得到第二个点是一个重复点，则可以推断出满足DCM模式的第三种，含有多个点，但是都是重复点，则继续解码重复点的数目是否大于1(熵解码)，如果大于1，则继续解码剩余重复点的数目(利用指数哥伦布进行解码)。
如果当前节点不满足DCM节点的要求，则直接退出(即点数大于2个点，并且不是重复点)。
在解码完成当前节点的点数目之后，对当前节点中包含点的坐标信息进行解码。下面将分别对激光雷达点云和面向人眼点云进行详细介绍。
(一)面向人眼点云。
(1)如果当前节点中仅仅只含有一个点，则会对点的三个维度方向的几何信息进行直接解码(Bypass coding)；
(2)如果当前节点中含有两个点，则会首先通过利用点的几何坐标得到优先解码的坐标轴dirextAxis，这里需要注意的是，目前比较的坐标轴只包含x和y轴，不包含z轴。假设当前节点的几何坐标为nodePos，则判断的方式如下：
dirextAxis＝！(nodePos[0]<nodePos[1])(9)
也就是会将节点坐标几何位置小的轴作为优先解码的坐标轴dirextAxis，其次按照如下方式首先对优先解码的坐标轴dirextAxis几何信息进行解码。假设优先解码的轴对应的待解码几何bit深度为nodeSizeLog2，并假设两个点的坐标分别为pointPos[0]和pointPos[1]。具体编码过程如下：

在解码完成优先解码的坐标轴dirextAxis之后，再继续对当前点的几何坐标进行直接解码。假设每个点的剩余编码bit深度为nodeSizeLog2，并假设点的坐标信息为pointPos，则具体解码过程如下：
(二)面向激光雷达点云。
如果当前节点中含有两个点，则会首先通过利用点的几何坐标得到优先解码的坐标轴dirextAxis，假设当前节点的几何坐标为nodePos，则判断的方式如下：
dirextAxis＝！(nodePos[0]<nodePos[1])(10)
也就是会将节点坐标几何位置小的轴作为优先解码的坐标轴dirextAxis，这里需要注意的是，目前比较的坐标轴只包含x轴和y轴，不包含z轴。其次按照如下方式首先对优先编码的坐标轴dirextAxis几何信息进行解码，假设优先解码的轴对应的代编码几何bit深度为nodeSizeLog2，并假设两个点的坐标分别为pointPos[0]和pointPos[1]。具体编码过程如下：
在解码完优先解码的坐标轴dirextAxis之后，再对当前点的几何坐标进行解码。
同样的首先利用当前节点的几何信息nodePos得到一个直接解码的主轴方向，其次利用已经完成解码的方向的几何信息来对另外一个维度的几何信息进行解码。同样假设直接解码的轴方向是directAxis，并且假设直接解码中的待解码bit深度为nodeSizeLog2,则解码方式如下：
这里需要注意的是，在这里会将directAxis方向的几何精度信息全部解码。
在解码完成directAxis坐标方向的所有精度之后，会首先计算当前节点的LaserIdx，即nodeLaserIdx；其次会利用节点的LaserIdx即nodeLaserIdx来对点的LaserIdx即pointLaserIdx进行预测解码，其中节点或者点的LaserIdx的计算方式跟编码端相同。最终对当前点的LaserIdx与节点的LaserIdx预测残差信息进行解码得到ResLaserIdx，则解码方式如下：
PointLaserIdx＝nodeLaserIdx+ResLaserIdx(11)
在解码完成当前点的LaserIdx之后，对当前点三个维度的几何信息利用激光雷达的采集参数进行预测解码。具体算法如下：
如图11所示，首先利用当前点对应的LaserIdx得到对应的水平方位角的预测值，即其次利用当前点对应的节点几何信息得到节点对应的水平方位角度其中，假设节点的几何坐标为nodePos，则水平方位角与节点几何信息之间的计算方式如下：
通过利用激光雷达的采集参数，可以得到每个Laser的旋转点数numPoints，即代表每个激光射线旋转一圈得到的点数，则可以利用每个Laser的旋转点数计算得到每个Laser的旋转角速度deltaPhi，计算方式如下：
进一步地，利用节点的水平方位角以及当前点对应的Laser前一个编码点的水平方位角计算得到当前点对应的水平方位角预测值即如图17A和图17B所示的水平方位角的预测值。计算方式如下：
这样，通过利用水平方位角的预测值以及当前节点的低平面水平方位角和高平面的水平方位角来对当前节点的几何信息进行预测解码。具体如下所示：

int context＝(angLel≥0&&angLeR≥0)||(angLel<0&&angLeR<0)？0:2；
int absAngleL＝abs(angLel)；
int absAngleR＝abs(angLeR)；
context+＝absAngleL>absAngleR？0:1；
context+＝maxAngle>minAngle《1？4:0。
在解码完成点的LaserIdx之后，会利用当前点所对应的LaserIdx对当前点的Z轴方向进行预测解码，即当前通过利用当前点的x和y信息计算得到雷达坐标系的深度信息radius，其次利用当前点的激光LaserIdx得到当前点的正切值以及垂直方向的偏移量，则可以得到当前点的Z轴方向的预测值即Z_pred。具体如下所示：
int tanTheta＝tanθlaserIdx；
int zOffset＝ZlaserIdx；
Z_pred＝radius×tanTheta-zOffset。
进一步地，利用解码得到的Z_res和Z_pred来重建恢复得到当前点Z轴方向的几何信息。
对于基于三角面片集(triangle soup，trisoup)的几何信息编码而言，在基于trisoup的几何信息编码框架中，同样也要先进行几何划分，但区别于基于二叉树/四叉树/八叉树的几何信息编码，该方法不需要将点云逐级划分到边长为1×1×1的单位立方体，而是划分到子块(block)边长为W时停止划分，基于每个block中点云的分布所形成的表面，得到该表面与block的十二条边所产生的至多十二个交点(vertex)。依次编码每个block的vertex坐标，生成二进制码流。
对于基于trisoup的点云几何信息重建而言，在解码端进行点云几何信息重建时，首先解码vertex坐标用于完成三角面片重建，该过程如图19A、图19B和图19C所示。其中，图19A所示的block中存在3个交点(v1，v2，v3)，利用这3个交点按照一定顺序所构成的三角面片集被称为triangle soup，即trisoup，如图19B所示。之后，在该三角面片集上进行采样，将得到的采样点作为该block内的重建点云，如图19C所示。
对于基于预测树的几何编码(Predictive geometry coding，PredGeomTree)而言，基于预测树的几何编码包括：首先对输入点云进行排序，目前采用的排序方法包括无序、莫顿序、方位角序和径向距离序。在编码端通过利用两种不同的方式建立预测树结构，其中包括：KD-Tree(高时延慢速模式)和低时延快速模式(利用激光雷达标定信息)。在利用激光雷达标定信息时，将每个点划分到不同的Laser上，按照不同的Laser建立预测树结构。接下来基于预测树的结构，遍历预测树中的每个节点，通过选取不同的预测模式对节点的几何位置信息进行预测得到预测残差，并且利用量化参数对几何预测残差进行量化。最终通过不断迭代，对预测树节点位置信息的预测残差、预测树结构以及量化参数等进行编码，生成二进制码流。
对于基于预测树的几何解码而言，解码端通过不断解析码流，重构预测树结构，其次通过解析得到每个预测节点的几何位置预测残差信息以及量化参数，并且对预测残差进行反量化，恢复得到每个节点的重构几何位置信息，最终完成解码端的几何重构。
在几何编码完成后，需要对几何信息进行重建。目前，属性编码主要针对颜色信息进行。首先，将颜色信息从RGB颜色空间转换到YUV颜色空间。然后，利用重建的几何信息对点云重新着色，使得未编码的属性信息与重建的几何信息对应起来。在颜色信息编码中，主要有两种变换方法，一是依赖于 LOD划分的基于距离的提升变换，二是直接进行RAHT变换，这两种方法都会将颜色信息从空间域转换到频域，通过变换得到高频系数和低频系数，最后对系数进行量化并编码，生成二进制码流，具体参见图4A和图4B所示。
进一步地，在利用几何信息来对属性信息进行预测时，可以利用莫顿码进行最近邻居搜索，点云中每点对应的莫顿码可以由该点的几何坐标得到。计算莫顿码的具体方法描述如下所示，对于每一个分量用d比特二进制数表示的三维坐标，其三个分量可以表示为：
其中，xl,yl,zl∈{0,1}分别是x，y，z的最高位(l＝1)到最低位(l＝d)对应的二进制数值。莫顿码M是对x，y，z从最高位开始，依次交叉排列xl,yl,zl到最低位，M的计算公式如下所示：
其中，ml′∈{0,1}分别是M的最高位(l′＝1)到最低位(l′＝3d)的值。在得到点云中每个点的莫顿码M后，将点云中的点按莫顿码由小到大的顺序进行排列，并将每个点的权重值w设为1。
还可以理解，对于G-PCC编解码框架而言，通用测试条件如下：
(1)测试条件共4种：
条件1：几何位置有限度有损、属性有损；
条件2：几何位置无损、属性有损；
条件3：几何位置无损、属性有限度有损；
条件4：几何位置无损、属性无损。
(2)通用测试序列包括Cat1A，Cat1B，Cat3-fused，Cat3-frame共四类，其中Cat2-frame点云只包含反射率属性信息，Cat1A、Cat1B点云只包含颜色属性信息，Cat3-fused点云同时包含颜色和反射率属性信息。
(3)技术路线：共2种，以几何压缩所采用的算法进行区分。
技术路线1：八叉树编码分支。
在编码端，将包围盒依次划分得到子立方体，对非空的(包含点云中的点)的子立方体继续进行划分，直到划分得到的叶子结点为1×1×1的单位立方体时停止划分，在几何无损编码情况下，需要对叶子节点中所包含的点数进行编码，最终完成几何八叉树的编码，生成二进制码流。
在解码端，解码端按照广度优先遍历的顺序，通过不断解析得到每个节点的占位码，并且依次不断划分节点，直至划分得到1×1×1的单位立方体时停止划分，在几何无损解码的情况下，需要解析得到每个叶子节点中包含的点数，最终恢复得到几何重构点云信息。
技术路线2：预测树编码分支。
在编码端，通过利用两种不同的方式建立预测树结构，其中包括：基于KD-Tree(高时延慢速模式)和利用激光雷达标定信息(低时延快速模式)，利用激光雷达标定信息，可以将每个点划分到不同的Laser上，按照不同的Laser建立预测树结构。接下来基于预测树的结构，遍历预测树中的每个节点，通过选取不同的预测模式对节点的几何位置信息进行预测得到预测残差，并且利用量化参数对几何预测残差进行量化。最终通过不断迭代，对预测树节点位置信息的预测残差、预测树结构以及量化参数等进行编码，生成二进制码流。
在解码端，解码端通过不断解析码流，重构预测树结构，其次通过解析得到每个预测节点的几何位置预测残差信息以及量化参数，并且对预测残差进行反量化，恢复得到每个节点的重构几何位置信息，最终完成解码端的几何重构。
还需要说明的是，在如图4A或图4B所示，目前G-PCC编码框架包含三种属性编码方法：预测变换(Predicting Transform，PT)、提升变换(Lifting Transform，LT)以及区域自适应分层变换(Region Adaptive Hierarchical Transform，RAHT)。其中，前两者是以LOD的生成顺序为依据对点云预测编码，RAHT则是依据八叉树的构建层级自下而上对属性信息进行自适应变换。下面将分别对这三种点云属性编码方法进行具体介绍。
(a)点云属性信息的预测编码。
目前G-PCC的属性预测模块采用一种基于分层(Level-of-details，LoDs)结构的最近邻属性预测编码方案，LOD的构造方法包括基于距离的LOD构造方案、基于固定采样率的LOD构造方案以及基于八叉树的LOD构造方案等。在基于距离阈值的LOD构造方案中，构造LOD之前首先对点云进行Morton排序，来保证相邻点之间具有较强的属性相关性。图20为一种基于距离的LOD构造过程的示意图。如图20所示，根据用户提前预设的L个曼哈顿(Manhattan)距离(dl)，l＝0,1,…L-1；将点云划分成L个不同的点云细节层(Rl)，l＝0,1,…L-1，其中(dl)l＝0,1,…L-1满足dl<dl-1。LOD的构造过程如下所述：
(1)首先将点云中所有点都标记为未访问过，建立一个集合V用来存储已经访问过的点集；(2)对于每一次迭代l，通过对点云中的点进行遍历，如果当前点已经被访问过，则忽略该点，否则计算当前点到点集V的最小距离D，如果D<dl，则忽略该点；否则将当前点标记为已访问并将当前点加入细化层Rl和点集V；(3)细节层次LODl中的点由细化层R0,R1,R2…Rl中的点构成；(4)不断重复上述步骤，直至所有的点都被标记为已访问。
在LOD的结构基础上，每个点的属性值通过利用同一层或更高一层LOD中点的属性重建值进行线性加权预测，其中参考预测邻居的最大数目由编码器高层语法元素决定。对于每个点的属性，在编码端利用率失真优化算法选取通过利用搜索到的N个最近邻点的属性进行加权预测或者选择单个最近邻点的属性进行预测，最后对选取的预测模式以及预测残差进行编码。
其中，N代表点i最近邻点集中预测点的数目，Pi代表点i的N个最近邻点的合，Dm代表了最近邻点m到当前点i的空间几何距离，Attrm代表了最近邻点m重建之后的属性值，Attri′代表了对当前点i的属性预测值，点数N为提前预设的数值。
为了权衡属性编码效率和不同LOD层之间的并行处理，在编码器高层语法元素引入了一个开关可以控制是否引入LOD层内预测。如果开启则启动LOD层内预测，可以利用同一LOD层内的点进行预测。需要注意的是，当LOD层的数目为1时，总是使用LOD层内预测。
图21为一种LOD生成过程的可视化结果示意图。如图21所示，这里提供了一种基于距离的LOD生成过程的主观示例。具体是(从左向右)：第一层中的点是代表点云的外轮廓；随着细节层的增加，点云细节描述逐渐清晰。
图22为一种属性预测的编码流程示意图。如图22所示，针对G-PCC属性预测的具体流程，对于原始点云，首先搜索第K个点的三个近邻点，然后进行属性预测；根据第K个点的属性预测值与第K个点的属性原始值进行差值计算，可以得到第K个点的预测残差；然后进行量化与算术编码，最终生成属性码率。
(i)最优预测值选取：
LOD构建完成以后，根据LOD的生成顺序，首先从已编码的数据点中找到当前待编码点的三个最近邻点。将这3个最近邻点的属性重建值，作为当前待编码点的候选预测值；然后，根据率失真优化(Rate-Distortion Optimal,RDO)从中选择最优的预测值。例如，当编码图20中点P2的属性值时，将最近邻居点P4属性值的预测变量索引设为1；将次近邻点P5和三近邻点P0的属性预测变量索引分别设为2和3；将点P0、P5和P4的加权平均值的预测变量索引设为0，如表1所示；最后，利用RDO选择最佳预测变量。其中加权平均的公式如下所示：
其中，表示近邻点j到当前点i的空间几何权重：
表示对当前点i的属性预测值，j表示3个近邻点的索引，代表了近邻点重建之后的属性值，xi,yi,zi是当前点i的几何位置坐标，xij,yij,zij为近邻点j的几何坐标。
示例性地，表1提供了一种属性编码的候选预测项样本示例。
表1
(ii)属性预测残差及量化：
通过上述预测得到当前点i的属性预测值(k为点云的总点数)。令(ai)i∈0…k-1为当前点的属性原始值，则属性残差(ri)i∈0…k-1记为：
进一步对预测残差进行量化：
其中，Qi表示当前点i的量化后的属性残差，Qs为量化步长(Quantization step,Qs)，可以由CTC规定的量化参数QP(Quantization Parameter,QP)计算得出。
(iii)编码端重建属性值：
编码端重建的目的是为了后续点的预测。在重建属性值之前要对残差进行反量化，记为反量化后的残差：
与预测值相加得到点i的重建值
在基于LOD划分的基础上进行属性最近邻查找时，目前存在两大类算法：帧内最近邻查找和帧间最近邻查找。其中，帧间的最近邻查找算法具体如下，帧内的最近邻查找可以分为层间最近邻查找和层内最近邻查找两种算法。
(i)帧内最近邻查找：
帧内最近邻查找分为层间最近邻查找和层内最近邻查找两种算法。LOD划分之后，类似一个金字塔结构，如图23所示。
在一种具体的实现方式中，对于层间最近邻查找，金字塔结构如图24所示。图25为一种层间最近邻查找的LOD构造过程示意图。如图25所示，基于几何信息划分得到不同的LOD层，得到LOD0、LOD1和LOD2，利用LOD0中的点去预测下一层LOD中点的属性在层间最近邻查找的过程中。
下面将对帧内最近邻查找的整个过程进行详细地介绍。
在整个LOD的划分过程中，存在三个集合O(k)、L(k)以及I(k)。其中，k为LOD划分时LOD层的索引，I(k)为当前LOD层划分时的输入点集，经过LOD划分，得到O(k)集合以及L(k)集合，O(k)集合存储的是采样点集，L(k)为当前LOD层中的点集。即整个LOD划分的过程如下：
(1)初始化。
if k＝0,L(k)←{}；否则，L(k)←L(k-1)；
O(k)←{}；
(2)利用LOD划分算法，将采样点存入O(k)，其余的点划分到L(k)；
(3)进行下一次迭代时I←O(k)。
这里需要注意的是，由于整个LOD划分的过程是基于莫顿码进行划分的，因此O(k)、L(k)以及I(k)存储的是点对应的莫顿码索引。
在进行层间最近邻查找时，即L(k)集合中的点在O(k)集合中进行最近邻查找，具体的查找算法如下：
以基于空间关系进行最近邻查找为例，在对当前点P进行预测时，通过利用点P对应的父块(Block B)进行邻居搜索，如图26所示，搜索与当前父块共面、共线邻居块内的点来进行属性预测。
其中，图27A示出了一种共面的空间关系示意图，这里共有6个与当前父块具有关系的空间块。图27B示出了一种共面和共线的空间关系示意图，这里共有18个与当前父块具有关系的空间块。图27C示出了一种共面、共线和共点的空间关系示意图，这里共有26个与当前父块具有关系的空间块。
首先，利用当前点的坐标得到对应的空间块，其次在之前已编码的LOD层中进行最近邻查找，查找与当前块共面、共线和共点的空间块来得到当前点的N近邻。
当进行共面、共线和共点最近邻查找之后，仍然没有得到当前点的N近邻，则会基于快速查找算法来得到当前点的N近邻，具体算法如下：
如图28所示，当进行属性层间预测时，首先利用当前待编码点的几何坐标得到当前点所对应的莫顿码，其次基于当前点的莫顿码在参考帧中查找到第一个大于当前点莫顿码的参考点(j)，其次在[j-searchRange,j+searchRange]范围内进行最近邻查找。
其余具体的更新最近邻的算法和帧间最近邻查找算法一致，在这里不在详述，具体的算法会在帧间最近邻查找算法中提到。
可以理解的是，在本申请实施例中，一个视频帧可以理解为一幅图像。示例性地，当前帧可以理解为当前图像，参考帧可以理解为参考图像。
在另一种具体的实现方式中，对于层内最近邻查找，图29示出了一种属性层内最近邻查找的LOD结构示意图。如图29所示，如果层内预测算法开启，即语法元素EnableRefferingSameLoD＝1，那么可以允许在层内最近邻查找，如对于LOD1层，当前点P6的最近邻点可以为P1，其他层不允许；如果语法元素EnableRefferingSameLoD＝0，那么允许在其他层进行层间查找，如对于LOD1层，当前点P6的 最近邻点可以为P4。也就是说，当层内预测算法开启时，会在同一层LOD内，在同层已编码的点集中进行最近邻查找，得到当前点的N近邻(同样进行层间最近邻查找)。
在进行属性层内预测时，会基于快速查找算法进行最近邻查找，具体的算法如如图30所示。其中，当前点用网格表示，假设当前点的莫顿码索引为i，则会在[i+1,i+searchRange]进行最近邻查找。具体的最近邻查找算法与帧间基于块的快速查找算法一致，在这里不再详述。
(ii)帧间最近邻查找：
图28为一种属性帧间预测示意图。如图28所示，当进行属性帧间预测时，首先利用当前待编码点的几何坐标得到当前点所对应的莫顿码，其次基于当前点的莫顿码在参考帧中查找到第一个大于当前点莫顿码的参考点(j)，其次在[j-searchRange,j+searchRange]范围内进行最近邻查找。
目前的帧内和帧间进行最近邻查找时，是基于块进行邻域查找的，具体的参见图31。如图31所示，在对当前点(莫顿码索引为i)进行邻域查找时，首先将参考帧中的点按照莫顿码划分成N(N＝3)个层，具体的划分算法如下：
第一层：将假设参考帧的点为numPoints，首先将参考帧中的点每M(M＝25＝32)个点划分到一个块
中；
第二层：在第一层的基础上，同样按照莫顿码的顺序对第一层的块每M(M＝25＝32)个块划分到一个
块中；
第三层：在第二层的基础上，同样按照莫顿码的顺序对第二层的块每M(M＝25＝32)个块划分到一个
块中；
最终得到如图31所示的预测结构。
在基于如图31所示的预测结构来进行属性预测，假设当前待编码点的莫顿码索引为i，首先在参考帧中得到第一个大于等于当前点莫顿码的点，索引为j。其次基于j计算得到参考点的块索引，具体计算方式如下：
第一层：BucketSize_0＝25＝32；
第二层：BucketSize_1＝25＝32×BucketSize_0＝1024；
第三层：BucketSize_2＝25＝32×BucketSize_1＝32768。
假设当前点的预测帧中的参考范围为[j-searchRange,j+searchRange]，利用j-searchRange计算得到第三层的起始索引，j+searchRange计算得到第三层的终止索引；其次，首先在第三层的块中判断第二层的一些块是否需要进行最近邻查找，其次到第二层，对于第一层中的每个块判断是否需要进行查找，如果第一层的某些块需要进行最近邻查找，则会对第一层中的一些块中点进行逐点判断来更新最近邻。
下面介绍一下，基于索引计算块的算法，假设当前点对应的莫顿码索引为index，那么对应的第三层块的索引为：
idx_2＝index/BucketSize_2(24)
在得到第三层的块索引idx_2之后，可以利用idx_2得到当前块在第二层对应的块的起始索引和终止索引：
startIdx1＝idx_2×BucketSize_1(25)
endIdx＝idx_2×BucketSize_1+BucketSize_1-1(26)
同样，基于同样的算法基于第二层块的索引得到第一层块的索引。
在基于块进行最近邻查找时，会首先判断当前块是否需要进行最近邻查找，也就是筛选块的最近邻查找。每个空间块可以通过两个变量进行得到minPos和maxPos，minPos表示的是块的最小值，maxPos表示的是块的最大值。
假设当前点查找的N近邻中最远点的距离为Dist，待编码点的坐标为(x,y,z)，当前块表示为(minPos,maxPos),其中minPos为包围盒三个维度上的最小值，maxPos为包围盒三个维度上的最大值，则当前点与包围盒之间的距离D计算如下：
int dx＝int(std::max(std::max(minPos[0]-point[0],0),point[0]-maxPos[0]))；
int dy＝int(std::max(std::max(minPos[1]-point[1],0),point[1]-maxPos[1]))；
int dz＝int(std::max(std::max(minPos[2]-point[2],0),point[2]-maxPos[2]))；
D＝dx+dy+dz；
当D小于等于Dist，才会去遍历当前块中的点。
(b)点云属性信息的提升变换编码。
图32为一种提升变换的编码流程示意图。提升变换同样是基于LOD对点云属性进行预测编码。与预测变换的不同之处在于，提升变换首先会对LOD进行高低层的划分，按照LOD生成层的逆序进行预测，并且在预测的过程中引入了更新算子来对低层LOD中点的量化权重进行更新，以提高预测的 准确性。这是由于低层LOD中点的属性值会频繁的用于高层LOD中点的属性值预测，低层LOD中的点应具有更大的影响力。
步骤1：分割过程。
分割过程是将完整的LOD层分为低LOD层L(N)和高LOD层H(N)。如果某点云有三层LOD，即(LODl)l＝0,1,2，经过分割后，LOD2为高LOD层，记为H(N)，(LODl)l＝0,1为低LOD层，记为L(N)。
步骤2：预测过程。
高层LOD中的点从低层中选取最近邻点的属性信息作为当前待编码点的属性预测值P(N)，预测残差D(N)记为：
D(N)＝H(N)-P(N)(27)
步骤3：更新过程。
对高层LOD中的属性预测残差D(N)进行更新，得到U(N)，并利用U(N)对低层LOD中点的属性值进行提升，如下式所示：
L′(N)＝L(N)+U(N)(28)
上述过程将依据LOD从高到低的顺序，不断迭代直至最低层LOD。
由于基于LOD的预测方案使得LOD低层中的点具有更大的影响力，基于提升小波变换的变换方案通过引入量化权重，并且根据预测残差D(N)以及预测点和相邻点之间的距离来更新预测残差，最后利用变换过程中的量化权重来对预测残差进行自适应量化。这里需要注意的是，在解码端可以通过几何重构来确定每个点的量化权重值，因此不要对量化权重进行编码。
(c)区域自适应分层变换。
区域自适应分层变换(RAHT)是一种哈尔小波变换，它可以将点云属性信息从空域变换到频域，进一步减少点云属性之间的相关性。其主要思想是按照八叉树结构，采用自底向上的方式对每一层中的节点分别从X、Y、Z三个维度进行变换(如图34)，并迭代直至八叉树的根节点。如图33所示，其基本思想是基于八叉树的层级结构进行小波变换，将属性信息与八叉树节点相关联，对于同一父节点中被占据节点的属性沿着自底向上的方式进行递归变换，对于每一层中的节点分别从X、Y、Z三个维度进行变换，直至变换至八叉树的根节点。在分层变换的过程中，将同层节点变换之后得到的低通/低频(DC)系数传递到下一层的节点继续进行变换，而所有的高通/高频(AC)系数可以通过算术编码器进行编码。
在变换过程中，同一层节点变换之后的DC系数(直流分量)将传递到上一层继续变换，而每一层变换后的AC系数(交流分量)将进行量化编码。下面将介绍主要的变换过程。
图35A为一种RAHT正变换的过程示意图，图35B为一种RAHT逆变换的过程示意图。针对RAHT对应的变换与逆变换过程，假设g′L,2x,y,z和g′L,2x+1,y,z为L层中互为近邻点的两个属性DC系数。经过线性变换后，L-1层的信息为AC系数f′L-1,x,y,z和DC系数g′L-1,x,y,z；然后，f′L-1,x,y,z将不再进行变换，直接进行量化编码，g′L-1,x,y,z将继续寻找近邻进行变换，如果寻找不到，则将其直接传递至L-2层，即RAHT变换仅对存在邻居点的节点有效，没有邻居点的节点将直接传递至上一层。在上述变换过程中，g′L,2x,y,z和g′L,2x+2,y,z对应的权重(该节点内非空子节点的个数)分别为w′L,2x,y,z和w′L,2x+1,y,z(简写为w′0和w′1)，g′L-1,x,y,z的权重为w′L-1,x,y,z，则通用变换公式为：
其中，Tw0,w1为变换矩阵：
变换矩阵会随着各点对应的权重自适应变化更新。上述过程会依据八叉树的划分结构不断迭代更新，直至八叉树的根节点。
在一种具体的实现方式中，针对区域自适应分层帧内预测变换编码，可以基于RAHT变换编码的基础上进行预测。如图33所示，RAHT属性变换基于八叉树层级的顺序，由体素级别不断进行变换直至得到根节点，从而完成整个属性的分层变换编码。在预测变换编码中，同样基于八叉树的层级顺序进行属性预测变换编码，但是是从根节点不断进行变换直至到体素级别。在每一次RAHT属性变换的过程中，是基于2×2×2的块进行属性预测变换编码。具体的如图36所示。如图36所示，可以看到网格填充块为当前待编码块，斜线填充块为与当前待编码块共面和共线的一些邻域块。其中，当前块的属性通过如下方式进行归一化处理：
Anode＝∑p∈nodeattribute(p)；
wnode＝∑p∈node 1＝{p∈node}；
anode＝Anode/wnode。
首先，可以通过当前块中包含点的属性得到当前块的属性，即：Anode。通过对当前块中包含点属性进行简单的相加，其次利用当前块的属性与的当前块中点的个数进行归一化处理得到当前块属性的均值anode。利用当前块属性的均值进行属性变换编码。具体编码过程参见图37。
如图37所示，这里示出了RAHT属性预测变换编码的整体流程。其中，(a)为当前块以及共面和共线的一些邻域块，(b)为经过归一化处理后的块，(c)为经过上采样后的块，(d)为当前块的属性，(e)为通过利用当前块的邻域属性进行线性加权拟合得到预测块的属性，最终将对两者分别进行属性变换，得到DC和AC系数，对AC系数进行预测编码。
其中，当前块的预测属性可以通过利用如图38所示进行线性拟合得到。如图38所示，首先得到当前块的19个邻域块，其次利用邻域块与当前块的每个子块之间的空间几何距离对每个子块的属性进行线性加权预测，最终利用线性加权得到的预测块属性进行变换。具体的属性变换如图39所示。
在图39中，(d)表示属性原始值，对应的属性变换系数如下：
(e)表示属性预测值，对应的属性变换系数如下：
根据属性原始值与属性预测值进行减法运算，可以得到预测残差如下：
在另一种具体的实现方式中，针对区域自适应分层帧间预测变换编码，在G-PCC属性帧间预测编码方案一中，类似帧内预测编码的过程。首先，基于几何信息构建RAHT属性变换编码结构，即：由体素级别不断进行变换直至得到根节点，从而完成整个属性的分层变换编码。按照这样的方式，构建得到帧内编码结构和帧间编码结构。其中，RAHT属性的帧间编码结构可以参见图40。
如图40所示，首先利用当前待编码节点的几何信息在参考帧中得到待编码节点的同位预测节点，其次利用参考节点的几何信息和属性信息得到当前待编码节点的预测属性。
其中，根据如下两种不同的方式得到当前待编码节点的属性预测值：
①当前节点的帧间预测节点有效：即同位节点存在，则将预测节点的属性直接作为当前待编码节点的属性预测值；
②当前节点的帧间预测节点无效：即同位节点不存在，则利用帧内相邻节点的属性预测值作为待编码节点的属性预测值。
最终，利用得到的属性预测值来对当前待编码节点的属性进行预测。从而完成整个属性的预测编码。
在又一种具体的实现方式中，针对区域自适应分层帧间预测变换编码，在G-PCC属性帧间预测编码方案二中，与帧内预测编码以及帧间预测编码方案一不同的是，如果启动帧间预测编码方案二，那么首先会基于当前待编码节点的几何信息构建RAHT属性变换编码结构，即由体素级别不断进行节点合并，直至得到整个RAHT变换树的根节点，从而完成整个属性的变换编码分层结构。其次，在根据RAHT变换结构，由根节点进行划分，得到每个节点的N个子节点(N小于等于8)，在帧间预测编码方案二中，会首先利用RAHT变换对N个子节点的属性进行独立正交变换，得到DC系数和AC系数，其次按照以下方式来对N个子节点的AC系数进行属性帧间预测：
①当前节点的帧间预测节点有效：即同位节点存在，则将预测节点的属性直接作为当前待编码节点的属性预测值
②当前节点可以在参考帧的缓存中查找到与当前节点位置完全相同的节点：即同位节点存在，则将同位节点中包含的M个子节点的AC系数直接作为当前节点N个子节点的AC系数属性预测值。
如果预测节点的AC系数不为零：则将预测节点的AC系数直接作为预测值；
如果预测节点的AC系数为零，则会将帧内预测对应子节点的AC系数作为预测值。
③当前节点的帧间预测节点无效：即同位节点不存在，则利用帧内相邻节点的属性预测值作为待编码节点的属性预测值。
简单来讲，在G-PCC的RAHT属性变换编码时，可以按照由根节点到子节点的顺序进行编解码。首先，利用当前层节点的几何信息，依次按照Z、Y和X的顺序进行恢复得到当前层的子节点，其次，利用父节点层已经重建的属性来对当前层节点的属性进行预测解码，从而恢复得到当前层节点的属性，直至变换到子节点，即体素级别。但是在G-PCC中，如果当前层节点的数目与当前层节点子节点的数目完全一致时，此时表明当前层中的每个节点仅仅只有一个子节点，即当前层不会产生AC系数，但是在已有的编码方案中，仍然需要对当前层的节点依次进行变换、预测等过程，这样会导致RAHT属性变换编解码的复杂度增大，引入了冗余操作。同样的，在已有的RAHT编码方案中，编解码端首先完成非节点层属性信息的编码以及解码(即节点的大小大于等于1×1×1)，最终在对体素级别节点的属性信息进行编码和解码，原因是由于点云中会存在一种情况，即点云中存在重复点，因此需要先完成非体素级别点的属性信息编码和解码，其次在完成体素级别点属性信息的编码和解码。由于在当前编码单元中不存在重复点时，仍然需要对体素级别点属性信息进行编码及解码过程，进一步提升了属性变换编码/解码的时间复杂度。
基于此，本申请实施例提供了一种编码方法，首先确定当前单元的体素节点的第一数量和当前单元的重建节点的第二数量；然后根据第一数量与第二数量，确定当前单元的体素节点的属性信息是否跳过编码。本申请实施例提供了还一种解码方法，首先确定当前单元的体素节点的第一数量和当前单元的重建节点的第二数量；其中，第一数量与第二数量用于确定是否对当前单元的体素节点进行跳过解码；然后根据第一数量与第二数量，确定当前单元的体素节点的属性重建值。
如此，在对每个体素节点进行属性重建时，优化了每个体素节点是否进行属性编解码的判断条件，具体为如果当前单元不存在重复节点，即第一数量与第二数量相同，此时不需要编解码当前单元的体素节点，从而在保证点云属性的编解码效率基础上，可以降低点云属性编解码的时间复杂度，并且还可以节省码率，进而提升了点云的编解码性能。
下面将结合附图对本申请各实施例进行详细说明。
在本申请的一实施例中，参见图41，其示出了本申请实施例提供的一种解码方法的流程示意图。如图41所示，该方法可以包括：
S4101：确定当前单元的体素节点的第一数量和当前单元的重建节点的第二数量；其中，第一数量与第二数量用于确定是否对当前单元的体素节点进行跳过解码。
需要说明的是，本申请实施例的解码方法应用于点云解码器(可简称为“解码器”)。其中，该方法可以是指点云解码方法，具体是一种点云属性解码方法，更具体地，可以是一种点云属性RAHT变换预测的跳过解码方法。在这里，主要是针对当前层的节点是否跳过解码以及在划分到体素级别时的体素节点是否跳过解码进行优化，从而可以降低属性变换编解码的时间复杂度，并且对属性变换的编解码效率不会产生任何影响。
还需要说明的是，在本申请实施例中，当前单元可以是当前待解码的当前解码单元，这里可以是一个片(slice)。在一些实施例中，该方法还可以包括：对当前单元中的节点进行划分，确定至少一层；在最后一层的节点被划分到体素级别时，确定当前单元的体素节点以及体素节点的第一数量。
示例性地，在RAHT属性变换中，RAHT属性变换的顺序是从根节点依次进行划分，直至划分到体素级别，具体为划分到1×1×1大小的单位立方体时停止划分，从而完成整个点云属性的编码和重建。在这里，如图42所示，每次沿着Z方向、Y方向和X方向做一次下采样所得到的层即为一个RAHT变换层，即layer。然后直至划分到1×1×1大小的单位立方体时表示已经划分到体素级别，这时候可以确定体素节点以及体素节点的第一数量。
还需要说明的是，在本申请实施例中，体素节点表示由根节点划分到体素级别时对应的节点，体素节点的大小为1×1×1；重建节点表示当前单元中进行属性重建的节点。其中，在当前单元中的节点进行属性解码之前，当前单元中的节点的几何信息已经全部解码完成，这时候可以确定当前单元的重建节点的第二数量，从而能够判断第一数量与第二数量是否一致。也就是说，在本申请实施例中，在对体素节点进行属性解码之前，首先需要判断第一数量与第二数量是否相同，进而来确定当前单元的体素节点是否进行跳过解码。
在一些实施例中，若第一数量与第二数量相同，则跳过解码当前单元的体素节点。
在一些实施例中，若第一数量与第二数量不同，则对体素节点中的重复节点进行属性解码，以及跳过解码体素节点中除重复节点之外的剩余体素节点。
在本申请实施例中，当前单元划分的体素节点中可能会存在重复节点，因此第一数量与第二数量有可能不一致。其中，如果当前单元划分的体素节点中不存在重复节点，那么第一数量与第二数量一致；反之，如果当前单元划分的体素节点中存在重复节点，那么第一数量与第二数量不一致，此时需要针对 重复节点进行属性解码。在这里，重复节点(也可简称为“重复点”)是指几何信息一样、但属性信息不同的多个节点。
S4102：根据第一数量与第二数量，确定当前单元的体素节点的属性重建值。
需要说明的是，在本申请实施例中，对于当前单元的体素节点的属性重建值，可以是直接复制当前单元的重建节点的属性重建值，但是其中的重复节点的属性重建值则无法复制得到，此时需要通过解码码流来确定。也就是说，这里可以是由第一数量与第二数量之间的大小来决定如何确定当前单元的体素节点的属性重建值。
在一些实施例中，在第一数量与第二数量相同时，根据第一数量与第二数量，确定当前单元的体素节点的属性重建值，可以包括：将当前单元的体素节点的属性重建值设置为当前单元的重建节点的属性重建值。
示例性地，如果第一数量与第二数量相同，那么可以跳过解码当前单元的体素节点的属性信息，这时候可以将体素节点的属性重建值直接复制为当前单元的重建节点的属性重建值。
在一些实施例中，在第一数量与第二数量不同时，根据第一数量与第二数量，确定当前单元的体素节点的属性重建值，可以包括：解码码流，确定重复节点的属性重建值；将重复节点的属性重建值作为当前单元中第一重建节点的属性重建值，以及将体素节点中除重复节点之外的剩余体素节点的属性重建值设置为当前单元中除第一重建节点之外的剩余重建节点的属性重建值。
示例性地，如果第一数量与第二数量不同，表明当前单元划分的体素节点中存在有重复节点，这时候可以解码确定重复节点的属性重建值，然后将重复节点的属性重建值作为当前单元中第一重建节点的属性重建值，以及将重复节点之外的剩余体素节点的属性重建值复制为当前单元中除第一重建节点之外的剩余重建节点的属性重建值。
示例性地，对于几何信息相同、属性信息不同的多个节点(即重复节点)，编码端可以采用计时器来实现把这多个节点的属性信息进行相加，将相加值确定为该几何信息对应节点的属性信息；后续在解码端，同样可以采用计时器来实现从这个属性信息中将这多个节点各自的属性信息分别剥离出来，以得到重复节点的属性重建值。
在一些实施例中，解码码流，确定重复节点的属性重建值，可以包括：设置计时器；在重复节点的属性重建值开始解码时，启动计时器，并在计时器的计时达到预设数值时，确定重复节点的属性重建值全部解码完成。
需要说明的是，在第一数量与第二数量不同时，还可以根据第一数量与第二数量之间的差值，确定体素节点中的重复节点的第三数量；其中，计时器的设置与第三数量有关联关系。
示例性地，可以采用一个计时器来确定重复节点的属性重建值是否全部解码完成。如果计时器处于正计数模式，那么在重复节点的属性重建值开始解码时，计时器的初始值为0，在计时器计数到第三数量时，此时重复节点的属性重建值全部解码完成；如果计时器处于倒计数模式，那么在重复节点的属性重建值开始解码时，计时器的初始值为第三数量，在计时器计数到0时，此时重复节点的属性重建值全部解码完成。在重复节点的属性重建值全部解码完成之后，意味着后续的体素节点中不存在重复节点，可以跳过这些节点的属性解码，并且将后续剩余体素节点的属性重建值直接复制为最终剩余重建节点的属性重建值。
可以理解地，在本申请实施例中，至少一层包括当前层，其中，当前层可以为当前待解码的RAHT变换层，或称为“RAHT属性解码层”。图43为本申请实施例提供的另一种解码方法的流程示意图。如图43所示，该方法可以包括：
S4301：确定当前层的节点的第四数量和当前层的节点对应的子节点的第五数量；其中，第四数量和第五数量用于确定是否对当前层进行跳过解码。
需要说明的是，在本申请实施例中，可以先确定当前层的节点的第四数量，同时可以确定当前层的节点对应的子节点的第五数量。其中，在当前层的节点进行属性解码之前，由于当前层的节点的几何信息已经解码完成，根据当前层的节点的几何信息可以确定出当前层的节点数目(即“第四数量”)和当前层的节点的子节点数目(即“第五数量”)。
进一步地，在本申请实施例中，需要先基于点云中的点的几何信息来构建RAHT属性变换结构，可以按照由根节点到子节点的顺序进行解码。利用当前层节点的几何信息，依次按照Z、Y和X的顺序进行恢复得到当前层的子节点，其次利用上一层的节点已经重建的属性来对当前层节点的属性进行预测解码，从而恢复得到当前层节点的属性重建值，直至变换到体素级别，进而可以得到包括有至少一个RAHT变换层的RAHT属性变换结构。
需要说明的是，在本申请实施例中，RAHT属性变换可以是基于八叉树层级的顺序执行的。其中，基于八叉树的层级顺序，可以由体素级别不断进行变换直至到根节点，从而构建出八叉树。然后在预测 变换过程中，同样基于八叉树的层级顺序进行属性预测变换编码，但是是由根节点不断进行变换直至到体素级别。
可以理解的是，在本申请实施例中，可以定义每次沿着预设方向，如Z方向、Y方向和X方向依次做一次下采样得到的层即为一个RAHT变换层，如当前层(layer)。
还需要说明的是，在本申请实施例中，对于当前层来说，当前层中可以包括至少一个点。其中，对于当前层中的至少一个点，在对当前层进行解码时，其可以作为当前层中的待解码节点。
进一步地，在本申请实施例中，对于当前层中的每一个点，其对应一个几何信息和一个属性信息；其中，几何信息表征该点的空间关系，属性信息表征该点的属性的相关信息。
在这里，属性信息可以为颜色信息，也可以是反射率或者其它属性，本申请实施例不作具体限定。其中，当属性信息为颜色信息时，具体可以为任意颜色空间的颜色信息。示例性地，属性信息可以为RGB空间的颜色信息，也可以为YUV空间的颜色信息，还可以为YCbCr空间的颜色信息等等，本申请实施例也不作具体限定。
还需要说明的是，在本申请实施例中，在进行RAHR变换编码时，首先是完成非体素级别的节点属性变换以及逆变换，其次完成体素级别的节点变换，因为点云中会存在一种情况，即点云中存在重复节点的现象。
相应地，在本申请实施例中，如果当前层的节点为非体素级别时，那么第四数量可以表征当前层的被占据节点的数量；第五数量则可以表征当前层的节点中被占据的子节点的数量。
相应地，在本申请实施例中，如果当前层的节点为体素级别时，那么第四数量可以表征当前层的被占据节点的数量；第五数量则可以表征待解码的节点的数量。
也就是说，在本申请实施例中，第四数量即为当前层的有效节点(即被占据的节点)的数量，而对于非体素级别的节点，第五数量为当前层的节点的有效子节点(即被占据的子节点)的数量，对于体素级别的节点，第五数量为待解码的节点的数量。
进一步地，在本申请实施例中，可以先确定当前层的节点的几何信息；然后再根据几何信息确定当前层的节点对应的子节点，以及第五数量。
需要说明的是，在本申请实施例中，对于当前层的当前节点来说，在利用当前节点的几何信息进行对应的子节点的确定时，可以选择利用当前节点的几何信息进行上采样，得到当前节点被占据的子节点(子节点个数为N，其中N的取值最大为8)。
示例性地，在一些实施例中，在对当前层的节点的属性信息进行编解码时，可以先得到当前层的节点的数量，即第四数量；同时，在利用当前层的节点的几何信息进行恢复得到当前层的节点的子节点之后，可以得到当前层的节点的子节点的数量，即第五数量。
S4302：根据第四数量和第五数量，确定当前层的节点对应的子节点的属性重建值。
需要说明的是，在本申请实施例中，在确定当前层的节点对应的第四数量和第五数量之后，可以进一步根据第四数量和第五数量，确定当前层的节点对应的子节点的属性重建值。
进一步地，在本申请实施例中，在确定当前层的节点对应的第四数量和第五数量之后，可以利用第四数量和第五数量进行是否跳过编码层(当前层)的属性信息的编解码。
可以理解的是，在本申请实施例中，由于RAHT变换仅对存在邻居点的节点有效，如果当前层的节点的数目与当前层的节点的子节点的数目完全一致时，则可以表明当前层中的每个节点仅仅只有一个子节点；在这种情况下，当前层不会产生AC系数(高频系数)，因此可以选择跳过对当前层的节点所依次进行的变换、预测等过程。
也就是说，在本申请实施例中，通过利用当前层的节点数目与子节点数目，可以自适应地决定当前层是否可以跳过编解码。其中，判定是否跳过对当前层的编解码处理的关键，在于当前层的节点数目与子节点数目是否相同，即第四数量和第五数量是否相同。
在一些实施例中，在根据第四数量和第五数量，确定当前层的节点对应的子节点的属性重建值时，该方法还可以包括：若第四数量和第五数量相同，则将当前层的节点的属性重建值确定为当前层的节点对应的子节点的属性重建值。
可以理解的是，在本申请实施例中，如果当前层的节点对应的第四数量和第五数量相同，即可以确定当前层的节点的数量与当前层的节点所对应的子节点的数量是相同的，那么，便可以认为对于当前层的每一个节点来说，均只对应有一个子节点。
相应地，在本申请实施例中，由于RAHT变换仅对存在邻居点的节点有效，在RAHT属性变换的过程中，如果当前层的每一个节点均仅对应有一个子节点，那么可以认为当前层不会产生AC系数，因此可以选择不对当前层的节点依次进行变换、预测等过程，即跳过对当前层的处理，此时，可以称为“跳过解码层”。
相应地，在本申请实施例中，如果确定当前层的节点对应的第四数量和第五数量相同，即确定当前层为跳过解码层，那么可以选择跳过对当前层的节点依次进行变换、预测等过程，而是可以直接将当前层的节点的属性重建值确定为当前层的节点对应的子节点的属性重建值。
也就是说，在本申请实施例中，在RAHT属性变换的过程中，如果当前层的每一个节点均仅对应有一个子节点，那么可以选择不再对当前层的节点依次进行变换、预测等过程，从而可以降低RAHT属性变换编解码的复杂度。
进一步地，在本申请实施例中，在根据第四数量和第五数量，确定当前层的节点对应的子节点的属性重建值时，若第四数量和第五数量相同，确定当前层为跳过解码层，那么可以选择跳过当前层至下一层，然后将下一层作为当前层，继续判断是否对下一层的节点进行跳过解码。
相应地，在本申请实施例中，对于当前层的节点对应的子节点来说，可以先确定子节点对应的下一层子节点的第六数量；然后根据第五数量和第六数量，确定子节点对应的下一层子节点的属性重建值。
可以理解的是，在本申请实施例中，如果当前层的节点对应的子节点为非体素级别的节点，那么第六数量可以为当前层的节点对应的子节点的下一层有效子节点(即下一层被占据的子节点)的数量，如果当前层的节点对应的子节点为体素级别的节点，那么第六数量可以为待解码的节点的数量。
需要说明的是，在本申请实施例中，在确定定第五数量和第六数量之后，可以利用定第五数量和第六数量进行是否跳过编码层(当前层的下一层)的属性信息的编解码。
示例性地，在一些实施例中，如果第五数量和第六数量相同，那么可以选择不对当前层的节点的子节点依次进行变换、预测等过程，即跳过对当前层的子节点的处理，而是可以直接将当前层的节点对应的子节点的属性重建值确定为当前层的节点对应的子节点的下一层子节点的属性重建值。
在一些实施例中，在根据第四数量和第五数量，确定当前层的节点对应的子节点的属性重建值时，该方法还可以包括：若当前层的节点对应的第四数量和第五数量不同，则根据当前层的节点确定当前层的节点对应的子节点的属性预测值；基于子节点的属性预测值进行RAHT变换，确定当前层的节点对应的高频系数的重建值和低频系数；基于高频系数的重建值和低频系数进行RAHT逆变换，确定子节点的属性重建值。
需要说明的是，在本申请实施例中，如果第四数量和第五数量不相同，即可以确定当前层的节点的数量与当前层的节点所对应的子节点的数量是不同的，此时，当前层依然会产生AC系数，因此可以继续对当前层的节点依次进行变换、预测等过程，而不会跳过对当前层的处理，此时，可以将当前层作为非跳过解码层。
进一步地，在一些实施例中，在根据当前层的节点确定当前层的节点对应的子节点的属性预测值时，可以包括：确定当前层的节点对应的相邻节点；根据相邻节点对应的属性重建值和相对距离参数，确定当前层的节点对应的子节点的属性预测值。
还需要说明的是，在本申请实施例中，相邻节点可以是指当前节点的邻域节点。其中，相邻节点对应的相对距离参数可以表征当前层的节点对应的子节点与对应的相邻节点的之间的空间几何距离。
示例性地，在本申请实施例中，对于当前层的当前节点，该当前节点包括子节点1和子节点2这两个子节点，当前节点与相邻节点之间的相对距离参数可以包括子节点1与相邻节点的之间的空间几何距离，还可以包括子节点2与相邻节点的之间的空间几何距离。
示例性地，在本申请实施例中，在根据当前层的节点确定当前层的节点对应的子节点的属性预测值时，对于当前层的当前节点，可以利用当前节点的邻域节点的重建属性(属性重建值)以及每个邻域节点距离当前节点的子节点的空间几何距离进行线性拟合，最终得到当前节点的每个子节点的属性预测值。
示例性地，在本申请实施例中，对于当前层的当前节点来说，可以先确定当前节点的19个相邻节点，其次利用相邻节点与当前节点的每个子节点之间的空间几何距离对每个子节点的属性进行线性加权预测，最终得到每个子节点的属性预测值。
进一步地，在一些实施例中，在基于子节点的属性预测值进行RAHT变换，确定当前层的节点对应的高频系数的重建值和低频系数时，可以包括：基于子节点的属性预测值进行RAHT变换，确定当前层的节点对应的高频系数的预测值和低频系数；根据高频系数的预测值确定当前层的节点对应的高频系数的重建值。
可以理解的是，在本申请实施例中，对于当前层的当前节点来说，在确定出当前节点的子节点对应的属性预测值之后，可以利用对应的子节点的属性预测值进行RAHT属性变换，从而可以得到相应的DC系数和AC系数，即得到当前节点相应的DC系数和AC系数。其中，DC系数即为低频系数，AC系数即为高频系数。
需要说明的是，在本申请实施例中，对于当前层的当前节点来说，利用子节点对应的属性预测值进行RAHT属性变换所获得的AC系数可以理解为当前节点对应的高频系数的预测值。
进一步地，在一些实施例中，在根据高频系数的预测值确定当前层的节点对应的高频系数的重建值时，可以包括：解码码流，确定当前层的节点对应的量化后的系数残差；根据高频系数的预测值和量化后的系数残差，确定当前层的节点对应的高频系数的重建值。
进一步地，在一些实施例中，在根据高频系数的预测值和量化后的系数残差，确定当前层的节点对应的高频系数的重建值时，可以包括：对量化后的系数残差进行反量化，确定当前层的节点对应的反量化残差值；根据当前层的节点对应的反量化残差值和当前层的节点对应的高频系数的预测值，确定当前层的节点对应的高频系数的重建值。
示例性的，在本申请实施例中，可以对当前层的节点对应的反量化残差值和当前层的节点对应的高频系数的预测值进行求和计算，进而可以获得当前层的节点对应的高频系数的重建值。
还需要说明的是，在本申请实施例中，在确定当前层的节点的低频系数和高频系数的重建值之后，便可以基于高频系数的重建值和低频系数进行RAHT逆变换，进而可以确定子节点的属性重建值。
示例性地，在本申请实施例中，假设g′L,2x,y,z和g′L,2x+1,y,z为L层中互为近邻点的两个属性DC系数。经过线性变换后，L-1层的信息为AC系数f′L-1,x,y,z和DC系数g′L-1,x,y,z；然后，f′L-1,x,y,z将不再进行变换，直接进行量化编码，g′L-1,x,y,z将继续寻找近邻进行变换，如果寻找不到，则将其直接传递至L-2层，即RAHT变换仅对存在邻居点的节点有效，没有邻居点的节点将直接传递至上一层。在该变换过程中，g′L,2x,y,z和g′L,2x+2,y,z对应的权重(该节点内非空子节点的个数)分别为w′L,2x,y,z和w′L,2x+1,y,z(简写为w′0和w′1)，g′L-1,x,y,z的权重为w′L-1,x,y,z，则通用变换公式为：
其中，Tw0,w1为变换矩阵，变换矩阵会随着各点对应的权重自适应变化更新。RAHT的正向变换(也可称为“RAHT正变换”)如前述的图35A所示。
示例性地，在本申请实施例中，根据所得到的当前节点的子节点的DC系数和AC系数进行RAHT的逆向变换，可以恢复得到当前节点的子节点的属性重建值。其中，RAHT的逆向变换(也可称为“RAHT反变换”、“RAHT逆变换”)如前述的图35B所示。
也就是说，在本申请实施例中，如果当前层的节点对应的第四数量和第五数量不相同，那么可以继续对当前层的节点依次进行变换、预测等过程。具体地，对于当前层的当前节点来说，可以利用当前节点的相邻节点的重建属性以及每个相邻节点距离当前节点的每个子节点的空间几何距离进行线性拟合，得到当前节点的每个子节点的预测属性；接着，利用每个子节点的预测属性进行RAHT属性变换得到相应的DC和AC系数，然后利用预测节点的AC系数(高频系数的预测值)和从码流中解析得到的AC系数(系数差值)，来恢复得到当前待解码节点(当前节点)的AC系数(高频系数的重建值)，最后便可以利用当前节点的AC系数(高频系数的重建值)和DC系数进行RAHT反变换，从而恢复得到当前节点的每个子节点的属性重建值。
进一步地，在本申请实施例中，如果当前层为非跳过编码层，那么可以对当前层的节点依次进行变换、预测等过程，确定出当前层的节点对应的子节点的属性重建值；然后针对当前层节点的子节点来说，可以先确定子节点对应的下一层子节点的第六数量；然后根据第五数量和第六数量，确定子节点对应的下一层子节点的属性重建值。
也就是说，在本申请实施例中，无论当前层是否为跳过编码层，即无论是否对当前层的节点依次进行变换、预测等过程；仍然需要不断重复上述步骤，依然确定其它层的节点数量，以及对应的子节点数量，然后再根据节点数量和对应的子节点数量来判断是否执行跳过解码的处理。
相应地，在本申请实施例中，不断重复步骤S4301至步骤S4302的方法，依次从RAHT变换的根节点不断重复直至到RAHT的叶子结点层的最后一个节点，从而完成整个RAHT变换的属性解码。
进一步地，在一些实施例中，该方法还可以包括：解码码流，确定预测模式标识信息；在预测模式标识信息指示当前单元启动跳过解码模式时，执行第一数量和第二数量的确定步骤，和/或，执行第四数量和第五数量的确定步骤。
在本申请实施例中，预测模式标识信息至少为以下高层语法元素中的其中之一：属性参数集(Attribute Parameter Set，APS)对应的语法元素和属性块头信息(Attribute Block Head，ABH)对应的语法元素。
在本申请实施例中，若预测模式标识信息的取值为第一值，则确定当前单元启动跳过解码模式；若预测模式标识信息的取值为第二值，则确定当前单元不启动跳过解码模式。下面针对当前单元的体素节点是否启动跳过解码模式和当前层的节点是否启动跳过解码模式分别进行说明。
在一种具体的实施例中，该方法还可以包括：解码码流，确定第一预测模式标识信息；在第一预测模式标识信息指示当前单元的体素节点启动跳过解码模式时，执行第一数量和第二数量的确定步骤。
需要说明的是，在本申请实施例中，若第一预测模式标识信息的取值为第一值，则确定当前单元的体素节点启动跳过解码模式；若第一预测模式标识信息的取值为第二值，则确定当前单元的体素节点不启动跳过解码模式。
还需要说明的是，在本申请实施例中，只有在当前单元的体素节点启动跳过解码模式时，这时候可以进一步确定第一数量和第二数量，然后根据第一数量和第二数量的大小来决定当前单元的体素节点是否跳过解码。具体地，如果两者一致，则跳过解码体素级别的节点属性，否则，会采用一个计时器，当该计时器的剩余重复节点数目为零时，则表示后续中的点不存在重复节点，同样可以跳过处理后续点的属性解码，并且将后续体素节点的属性重建值直接复制为最终剩余重建点的属性重建值。
在另一种具体的实施例中，该方法还可以包括：解码码流，确定第二预测模式标识信息；在第二预测模式标识信息指示当前层的节点启动跳过解码模式时，执行第四数量和第五数量的确定步骤。
需要说明的是，在本申请实施例中，若第二预测模式标识信息的取值为第一值，则确定当前层的节点启动跳过解码模式；若第二预测模式标识信息的取值为第二值，则确定当前层的节点不启动跳过解码模式。
还需要说明的是，在本申请实施例中，只有在当前层的节点启动跳过解码模式时，这时候可以进一步确定第四数量和第五数量，然后根据第四数量和第五数量的大小来决定当前层的节点是否跳过解码，即当前层是否为跳过解码层。具体地，如果两者一致，则确定当前层为跳过解码层，这时候不再需要解码当前层的节点对应的子节点的属性重建值；如果两者不一致，则确定当前层不属于跳过解码层，这时候可以按照相关技术的解码方式进行RAHT预测以及解码。
还需要说明的是，在本申请实施例中，第一值与第二值不同，而且第一值和第二值可以是参数形式，也可以是数字形式。具体地，第一预测模式标识信息和第二预测模式标识信息可以是写入在概述(profile)中的参数，也可以是一个标志(flag)的取值，这里不作具体限定。另外，对于第一值和第二值而言，第一值可以设置为1，第二值可以设置为0；或者，第一值可以设置为0，第二值可以设置为1；或者，第一值可以设置为true，第二值可以设置为false；或者，第一值可以设置为false，第二值可以设置为true。其中，在本申请实施例中，第一值设置为1，第二值设置为0，但是不作具体限定。
示例性地，在一些实施例中，以第一值设置为1，第二值设置为0为例，解码码流，确定第二预测模式标识信息的取值；如果第二预测模式标识信息的取值为1，那么即可确定当前层的节点启动跳过解码模式，进而可以根据上述方法进一步确定当前层的节点对应的第四数量和第五数量；如果第二预测模式标识信息的取值为0，那么即可确定当前层的节点不启动跳过解码模式，则可以按照常见的帧内预测的方式或帧间预测的方式对当前层的节点进行属性解码处理。
也就是说，在本申请实施例中，如果解码码流确定的第一预测模式标识信息的取值为第一值，即确定当前单元的体素节点启动跳过解码模式，那么可以执行第一数量和第二数量的确定步骤，即执行图41所示的解码流程；如果解码码流确定的第二预测模式标识信息的取值为第一值，即确定当前层的节点启动跳过解码模式，那么可以执行第四数量和第五数量的确定流程，即执行图43所示的解码流程。
综上所述，在本申请实施例中，在对属性信息进行编解码时，如果当前层的节点数目与当前层的子节点数目一致，则认为当前层属于跳过编码层，因此不需要对当前层进行变换、预测、编码以及解码等过程，从而可以降低属性变换编解码的时间复杂度，并且对属性的编码效率不会产生任何影响。
示例性的，在标准文本的实现中，一种具体的SPEC修改如下所示：


也就是说，本申请实施例提出的解码方法，在对属性进行RAHT解码时，在每一个RAHT变换层通过判断当前层的节点数目与当前层的子节点数目是否一致，来决定当前RAHT变换层是否属于跳过编码层，即确定是否跳过对当前层的变换、预测等过程。如果当前层的节点数目与当前层的子节点数目一致，则认为当前层属于跳过编码层，不需要进行变换、预测、编码以及解码等过程，从而可以降低属性变换编码\解码的时间复杂度，并且对属性的编码效率不会产生任何影响。
本实施例提供了一种解码方法，这里提出了一种跳过当前层的解码方式，通过利用当前层的节点数目与子节点数目来自适应地决定当前层是否可以跳过解码，这样可以在保证编解码效率不变的前提下，降低编解码的时间复杂度；另外，这里还提出了一种跳过体素级别节点的解码方式，首先可以得到体素节点数目和重建节点数目，当重建节点数目与体素节点数目一致时，则认为当前单元中不存在重复节点，同样可以跳过解码体素级别的节点属性；如此，在保证点云属性的编解码效率基础上，可以降低点云属性编解码的时间复杂度，并且还可以节省码率，进而提升了点云的编解码性能。
在本申请的一实施例中，参见图44，其示出了本申请实施例提供的一种编码方法的流程示意图。如图44所示，该方法可以包括：
S4401：确定当前单元的体素节点的第一数量和所述当前单元的重建节点的第二数量。
需要说明的是，本申请实施例的编码方法应用于点云编码器(可简称为“编码器”)。其中，该方法可以是指点云编码方法，具体是一种点云属性编码方法，更具体地，可以是一种点云属性RAHT变换预测的跳过编码方法。在这里，主要是针对当前层的节点是否跳过编码以及在划分到体素级别时的体素节点是否跳过编码进行优化，从而可以降低属性变换编解码的时间复杂度，并且对属性变换的编解码效率不会产生任何影响。
还需要说明的是，在本申请实施例中，当前单元可以是当前待编码的当前编码单元，这里可以是一个片(slice)。在一些实施例中，该方法还可以包括：对当前单元中的节点进行划分，确定至少一层；在最后一层的节点被划分到体素级别时，确定当前单元的体素节点以及体素节点的第一数量。
示例性地，在RAHT属性变换中，RAHT属性变换的顺序是从根节点依次进行划分，直至划分到体素级别，具体为划分到1×1×1大小的单位立方体时停止划分，从而完成整个点云属性的编码和重建。在这里，如图42所示，每次沿着Z方向、Y方向和X方向做一次下采样所得到的层即为一个RAHT变换层，即layer。然后直至划分到1×1×1大小的单位立方体时表示已经划分到体素级别，这时候可以确定体素节点以及体素节点的第一数量。
还需要说明的是，在本申请实施，例中，体素节点表示由根节点划分到体素级别时对应的节点，体素节点的大小为1×1×1；重建节点表示当前单元中进行属性重建的节点。其中，在当前单元中的节点进 行属性编码之前，当前单元中的节点的几何信息已经全部编码完成，这时候可以确定当前单元的重建节点的第二数量，从而能够判断第一数量与第二数量是否一致。
S4402：根据第一数量与第二数量，确定当前单元的体素节点的属性信息是否跳过编码。
需要说明的是，在本申请实施例中，在对体素节点进行属性编码之前，首先需要判断第一数量与第二数量是否相同，进而来确定当前单元的体素节点是否进行跳过编码。
在一些实施例中，若第一数量与第二数量相同，则跳过编码当前单元的体素节点的属性信息。
在一些实施例中，若第一数量与第二数量不同，则对体素节点中的重复节点进行属性编码，以及跳过编码体素节点中除重复节点之外剩余体素节点的属性信息。
在本申请实施例中，当前单元划分的体素节点中可能会存在重复节点，因此第一数量与第二数量有可能不一致。其中，如果当前单元划分的体素节点中不存在重复节点，那么第一数量与第二数量一致；反之，如果当前单元划分的体素节点中存在重复节点，那么第一数量与第二数量不一致，此时需要针对重复节点进行属性编码。在这里，重复节点(也可简称为“重复点”)是指几何信息一样、但属性信息不同的多个节点。
在一些实施例中，对体素节点中的重复节点进行属性编码，可以包括：对重复节点的属性信息进行编码处理，将所得到的编码比特写入码流。
需要说明的是，在本申请实施例中，在对重复节点的属性信息进行编码处理时，可以包括：设置计时器；在重复节点的属性信息开始编码时，启动计时器，并在计时器的计时达到预设数值时，确定重复节点的属性信息全部编码完成。
还需要说明的是，在本申请实施例中，在第一数量与第二数量不同时，还可以根据第一数量与第二数量之间的差值，确定体素节点中的重复节点的第三数量；其中，计时器的设置与第三数量有关联关系。
示例性地，可以采用一个计时器来确定重复节点的属性重建值是否全部编码完成。如果计时器处于正计数模式，那么在重复节点的属性重建值开始编码时，计时器的初始值为0，在计时器计数到第三数量时，此时重复节点的属性重建值全部编码完成；如果计时器处于倒计数模式，那么在重复节点的属性重建值开始编码时，计时器的初始值为第三数量，在计时器计数到0时，此时重复节点的属性重建值全部编码完成。在重复节点的属性重建值全部编码完成之后，意味着后续的体素节点中不存在重复节点，可以跳过这些节点的属性编码。
进一步地，在一些实施例中，参见图45，在步骤S4401之后，该方法还可以包括：
S4501：根据第一数量与第二数量，确定当前单元的体素节点的属性重建值。
需要说明的是，在本申请实施例中，对于当前单元的体素节点的属性重建值，可以是直接复制当前单元的重建节点的属性重建值，但是其中的重复节点的属性重建值则无法复制得到。也就是说，这里可以是由第一数量与第二数量之间的大小来决定如何确定当前单元的体素节点的属性重建值。
在一些实施例中，在第一数量与第二数量相同时，根据第一数量与第二数量，确定当前单元的体素节点的属性重建值，可以包括：将当前单元的体素节点的属性重建值设置为当前单元的重建节点的属性重建值。示例性地，如果第一数量与第二数量相同，那么可以跳过编码当前单元的体素节点的属性信息，这时候可以将体素节点的属性重建值直接复制为当前单元的重建节点的属性重建值。
在一些实施例中，在第一数量与第二数量不同时，根据第一数量与第二数量，确定当前单元的体素节点的属性重建值，可以包括：将重复节点的属性重建值作为当前单元中第一重建节点的属性重建值，以及将体素节点中除重复节点之外的剩余体素节点的属性重建值设置为当前单元中除第一重建节点之外的剩余重建节点的属性重建值。
示例性地，如果第一数量与第二数量不同，表明当前单元划分的体素节点中存在有重复节点，这时候可以确定重复节点的属性重建值，然后将重复节点的属性重建值作为当前单元中第一重建节点的属性重建值，以及将重复节点之外的剩余体素节点的属性重建值复制为当前单元中除第一重建节点之外的剩余重建节点的属性重建值。另外，在第一数量与第二数量不同时，需要编码重复节点的属性信息，以便解码端可以解码确定这些重复节点的属性重建值。
示例性地，对于几何信息相同、属性信息不同的多个节点(即重复节点)，编码端可以采用计时器来实现把这多个节点的属性信息进行相加，将相加值确定为该几何信息对应节点的属性信息；后续在解码端进行属性解码时，采用计时器来实现从这个属性信息中将这多个节点各自的属性信息分别剥离出来，以得到重复节点的属性重建值。
还可以理解地，在本申请实施例中，至少一层包括当前层，其中，当前层可以为当前待解码的RAHT变换层，或称为“RAHT属性编码层”。图46为本申请实施例提供的又一种编码方法的流程示意图。如图46所示，该方法可以包括：
S4601：确定当前层的节点的第四数量和当前层的节点对应的子节点的第五数量；其中，第四数量 和第五数量用于确定是否对当前层进行跳过编码。
需要说明的是，在本申请实施例中，可以先确定当前层的节点的第四数量，同时可以确定当前层的节点对应的子节点的第五数量。其中，在当前层的节点进行属性编码之前，由于当前层的节点的几何信息已经编码完成，根据当前层的节点的几何信息可以确定出当前层的节点数目(即“第四数量”)和当前层的节点的子节点数目(即“第五数量”)。
进一步地，在本申请实施例中，需要先基于点云中的点的几何信息来构建RAHT属性变换结构，可以按照由根节点到子节点的顺序进行编码。利用当前层节点的几何信息，依次按照Z、Y和X的顺序进行恢复得到当前层的子节点，其次利用上一层的节点已经重建的属性来对当前层节点的属性进行预测编码，从而恢复得到当前层节点的属性重建值，直至变换到体素级别，进而可以得到包括有至少一个RAHT变换层的RAHT属性变换结构。
需要说明的是，在本申请实施例中，RAHT属性变换可以是基于八叉树层级的顺序执行的。其中，基于八叉树的层级顺序，可以由体素级别不断进行变换直至到根节点，从而构建出八叉树。然后在预测变换过程中，同样基于八叉树的层级顺序进行属性预测变换编码，但是是由根节点不断进行变换直至到体素级别。
可以理解的是，在本申请实施例中，可以定义每次沿着预设方向，如Z方向、Y方向和X方向依次做一次下采样得到的层即为一个RAHT变换层，如当前层(layer)。
还需要说明的是，在本申请实施例中，对于当前层来说，当前层中可以包括至少一个点。其中，对于当前层中的至少一个点，在对当前层进行编码时，其可以作为当前层中的待编码节点。
进一步地，在本申请实施例中，对于当前层中的每一个点，其对应一个几何信息和一个属性信息；其中，几何信息表征该点的空间关系，属性信息表征该点的属性的相关信息。
在这里，属性信息可以为颜色信息，也可以是反射率或者其它属性，本申请实施例不作具体限定。其中，当属性信息为颜色信息时，具体可以为任意颜色空间的颜色信息。示例性地，属性信息可以为RGB空间的颜色信息，也可以为YUV空间的颜色信息，还可以为YCbCr空间的颜色信息等等，本申请实施例也不作具体限定。
还需要说明的是，在本申请实施例中，在进行RAHR变换编码时，首先是完成非体素级别的节点属性变换以及逆变换，其次完成体素级别的节点变换，因为点云中会存在一种情况，即点云中存在重复节点的现象。
相应地，在本申请实施例中，如果当前层的节点为非体素级别时，那么第四数量可以表征当前层的被占据节点的数量；第五数量则可以表征当前层的节点中被占据的子节点的数量。
相应地，在本申请实施例中，如果当前层的节点为体素级别时，那么第四数量可以表征当前层的被占据节点的数量；第五数量则可以表征待编码的节点的数量。
也就是说，在本申请实施例中，第四数量即为当前层的有效节点(即被占据的节点)的数量，而对于非体素级别的节点，第五数量为当前层的节点的有效子节点(即被占据的子节点)的数量，对于体素级别的节点，第五数量为待编码的节点的数量。
进一步地，在本申请实施例中，可以先确定当前层的节点的几何信息；然后再根据几何信息确定当前层的节点对应的子节点，以及第五数量。
需要说明的是，在本申请实施例中，对于当前层的当前节点来说，在利用当前节点的几何信息进行对应的子节点的确定时，可以选择利用当前节点的几何信息进行上采样，得到当前节点被占据的子节点(子节点个数为N，其中N的取值最大为8)。
示例性地，在一些实施例中，在对当前层的节点的属性信息进行编解码时，可以先得到当前层的节点的数量，即第四数量；同时，在利用当前层的节点的几何信息进行恢复得到当前层的节点的子节点之后，可以得到当前层的节点的子节点的数量，即第五数量。
S4602：根据第四数量和第五数量，确定当前层的节点对应的子节点的属性重建值。
需要说明的是，在本申请实施例中，在确定当前层的节点对应的第四数量和第五数量之后，可以进一步根据第四数量和第五数量，确定当前层的节点对应的子节点的属性重建值。
进一步地，在本申请实施例中，在确定当前层的节点对应的第四数量和第五数量之后，可以利用第四数量和第五数量进行是否跳过编码层(当前层)的属性信息的编解码。
可以理解的是，在本申请实施例中，由于RAHT变换仅对存在邻居点的节点有效，如果当前层的节点的数目与当前层的节点的子节点的数目完全一致时，则可以表明当前层中的每个节点仅仅只有一个子节点；在这种情况下，当前层不会产生AC系数(高频系数)，因此可以选择跳过对当前层的节点所依次进行的变换、预测等过程。
也就是说，在本申请实施例中，通过利用当前层的节点数目与子节点数目，可以自适应地决定当前 层是否可以跳过编解码。其中，判定是否跳过对当前层的编解码处理的关键，在于当前层的节点数目与子节点数目是否相同，即第四数量和第五数量是否相同。
在一些实施例中，在根据第四数量和第五数量，确定当前层的节点对应的子节点的属性重建值时，该方法还可以包括：若第四数量和第五数量相同，则将当前层的节点的属性重建值确定为当前层的节点对应的子节点的属性重建值。
可以理解的是，在本申请实施例中，如果当前层的节点对应的第四数量和第五数量相同，即可以确定当前层的节点的数量与当前层的节点所对应的子节点的数量是相同的，那么，便可以认为对于当前层的每一个节点来说，均只对应有一个子节点。
相应地，在本申请实施例中，由于RAHT变换仅对存在邻居点的节点有效，在RAHT属性变换的过程中，如果当前层的每一个节点均仅对应有一个子节点，那么可以认为当前层不会产生AC系数，因此可以选择不对当前层的节点依次进行变换、预测等过程，即跳过对当前层的处理，此时，可以称为“跳过编码层”。
相应地，在本申请实施例中，如果确定当前层的节点对应的第四数量和第五数量相同，即确定当前层为跳过编码层，那么可以选择跳过对当前层的节点依次进行变换、预测等过程，而是可以直接将当前层的节点的属性重建值确定为当前层的节点对应的子节点的属性重建值。
也就是说，在本申请实施例中，在RAHT属性变换的过程中，如果当前层的每一个节点均仅对应有一个子节点，那么可以选择不再对当前层的节点依次进行变换、预测等过程，从而可以降低RAHT属性变换编解码的复杂度。
进一步地，在本申请实施例中，在根据第四数量和第五数量，确定当前层的节点对应的子节点的属性重建值时，若第四数量和第五数量相同，确定当前层为跳过编码层，那么可以选择跳过当前层至下一层，然后将下一层作为当前层，继续判断是否对下一层的节点进行跳过编码。
相应地，在本申请实施例中，对于当前层的节点对应的子节点来说，可以先确定子节点对应的下一层子节点的第六数量；然后根据第五数量和第六数量，确定子节点对应的下一层子节点的属性重建值。
可以理解的是，在本申请实施例中，如果当前层的节点对应的子节点为非体素级别的节点，那么第六数量可以为当前层的节点对应的子节点的下一层有效子节点(即下一层被占据的子节点)的数量，如果当前层的节点对应的子节点为体素级别的节点，那么第六数量可以为待编码的节点的数量。
需要说明的是，在本申请实施例中，在确定定第五数量和第六数量之后，可以利用定第五数量和第六数量进行是否跳过编码层(当前层的下一层)的属性信息的编解码。
示例性地，在一些实施例中，如果第五数量和第六数量相同，那么可以选择不对当前层的节点的子节点依次进行变换、预测等过程，即跳过对当前层的子节点的处理，而是可以直接将当前层的节点对应的子节点的属性重建值确定为当前层的节点对应的子节点的下一层子节点的属性重建值。
在一些实施例中，在根据第四数量和第五数量，确定当前层的节点对应的子节点的属性重建值时，该方法还可以包括：若当前层的节点对应的第四数量和第五数量不同，则根据当前层的节点确定当前层的节点对应的子节点的属性预测值；基于子节点的属性预测值进行RAHT变换，确定当前层的节点对应的高频系数的重建值和低频系数；基于高频系数的重建值和低频系数进行RAHT逆变换，确定子节点的属性重建值。
需要说明的是，在本申请实施例中，如果第四数量和第五数量不相同，即可以确定当前层的节点的数量与当前层的节点所对应的子节点的数量是不同的，此时，当前层依然会产生AC系数，因此可以继续对当前层的节点依次进行变换、预测等过程，而不会跳过对当前层的处理，此时，可以将当前层作为非跳过编码层。
进一步地，在一些实施例中，在根据当前层的节点确定当前层的节点对应的子节点的属性预测值时，可以包括：确定当前层的节点对应的相邻节点；根据相邻节点对应的属性重建值和相对距离参数，确定当前层的节点对应的子节点的属性预测值。
还需要说明的是，在本申请实施例中，相邻节点可以是指当前节点的邻域节点。其中，相邻节点对应的相对距离参数可以表征当前层的节点对应的子节点与对应的相邻节点的之间的空间几何距离。
示例性地，在本申请实施例中，对于当前层的当前节点，该当前节点包括子节点1和子节点2这两个子节点，当前节点与相邻节点之间的相对距离参数可以包括子节点1与相邻节点的之间的空间几何距离，还可以包括子节点2与相邻节点的之间的空间几何距离。
示例性地，在本申请实施例中，在根据当前层的节点确定当前层的节点对应的子节点的属性预测值时，对于当前层的当前节点，可以利用当前节点的邻域节点的重建属性(属性重建值)以及每个邻域节点距离当前节点的子节点的空间几何距离进行线性拟合，最终得到当前节点的每个子节点的属性预测值。
示例性地，在本申请实施例中，对于当前层的当前节点来说，可以先确定当前节点的19个相邻节 点，其次利用相邻节点与当前节点的每个子节点之间的空间几何距离对每个子节点的属性进行线性加权预测，最终得到每个子节点的属性预测值。
进一步地，在一些实施例中，在基于子节点的属性预测值进行RAHT变换，确定当前层的节点对应的高频系数的重建值和低频系数时，可以包括：基于子节点的属性预测值进行RAHT变换，确定当前层的节点对应的高频系数的预测值和低频系数；基于子节点的属性值进行RAHT变换，确定当前层的节点对应的高频系数和低频系数；根据当前层的节点对应的高频系数的预测值和当前层的节点对应的高频系数，确定当前层的节点对应的高频系数的重建值。
进一步地，在一些实施例中，在根据当前层的节点对应的高频系数的预测值和当前层的节点对应的高频系数，确定当前层的节点对应的高频系数的重建值时，可以包括：根据当前层的节点对应的高频系数的预测值和当前层的节点对应的高频系数，确定当前层的节点对应的系数残差；根据当前层的节点对应的系数残差，确定当前层的节点对应的高频系数的重建值。
进一步地，在一些实施例中，在根据当前层的节点对应的系数残差，确定当前层的节点对应的高频系数的重建值时，可以包括：对量化后的系数残差进行反量化，确定当前层的节点对应的反量化残差值；根据当前层的节点对应的反量化残差值和当前层的节点对应的高频系数的预测值，确定当前层的节点对应的高频系数的重建值。
需要说明的是，在本申请实施例中，对于当前层的当前节点来说，在确定出当前节点的子节点对应的属性预测值之后，可以利用对应的子节点的属性预测值进行RAHT属性变换，从而可以得到相应的DC系数和AC系数。其中，DC系数即为低频系数，AC系数即为高频系数。其中，在本申请实施例中，对于当前层的当前节点来说，利用子节点对应的属性预测值进行RAHT属性变换所获得的AC系数可以理解为当前节点对应的AC系数的预测值。
还需要说明的是，在本申请实施例中，对于当前层的当前节点来说，还可以利用子节点的属性值进行RAHT变换，确定当前层的节点对应的AC系数和DC系数。其中，在本申请实施例中，对于当前层的当前节点来说，利用子节点对应的属性值进行RAHT属性变换所获得的AC系数可以理解为当前节点对应的AC系数的原始值。
这样，根据当前层的节点对应的AC系数的预测值和AC系数的原始值，可以确定当前层的节点对应的系数残差；然后对量化后的系数残差进行反量化，确定当前层的节点对应的反量化残差值；再根据当前层的节点对应的反量化残差值和当前层的节点对应的AC系数的预测值，可以确定当前层的节点对应的AC系数的重建值。
示例性的，在本申请实施例中，可以对当前层的节点对应的反量化残差值和当前层的节点对应的AC系数的预测值进行求和计算，进而可以获得当前层的节点对应的AC系数的重建值。
进一步地，在一些实施例中，该方法还可以包括：对系数残差进行量化，确定当前层的节点对应的量化后的系数残差；对量化后的系数残差进行编码处理，将所得到的编码比特写入码流。
需要说明的是，在本申请实施例中，将量化后的系数残差写入码流，后续在解码端通过解码码流即可获得量化后的系数残差，然后根据高频系数的预测值和量化后的系数残差，就可以确定当前层的节点对应的高频系数的重建值。
还需要说明的是，在本申请实施例中，在确定当前层的节点的低频系数和高频系数的重建值之后，便可以基于高频系数的重建值和低频系数进行RAHT逆变换，进而可以确定子节点的属性重建值。
示例性地，在本申请实施例中，假设g′L,2x,y,z和g′L,2x+1,y,z为L层中互为近邻点的两个属性DC系数。经过线性变换后，L-1层的信息为AC系数f′L-1,x,y,z和DC系数g′L-1,x,y,z；然后，f′L-1,x,y,z将不再进行变换，直接进行量化编码，g′L-1,x,y,z将继续寻找近邻进行变换，如果寻找不到，则将其直接传递至L-2层，即RAHT变换仅对存在邻居点的节点有效，没有邻居点的节点将直接传递至上一层。在该变换过程中，g′L,2x,y,z和g′L,2x+2,y,z对应的权重(该节点内非空子节点的个数)分别为w′L,2x,y,z和w′L,2x+1,y,z(简写为w′0和w′1)，g′L-1,x,y,z的权重为w′L-1,x,y,z，则通用变换公式为：
其中，Tw0,w1为变换矩阵，变换矩阵会随着各点对应的权重自适应变化更新。RAHT的正向变换(也可称为“RAHT正变换”)如前述的图35A所示。
示例性地，在本申请实施例中，根据所得到的当前节点的子节点的DC系数和AC系数进行RAHT的逆向变换，可以恢复得到当前节点的子节点的属性重建值。其中，RAHT的逆向变换(也可称为“RAHT反变换”、“RAHT逆变换”)如前述的图35B所示。
也就是说，在本申请实施例中，如果当前层的节点对应的第四数量和第五数量不相同，那么可以继续对当前层的节点依次进行变换、预测等过程。具体地，对于当前层的当前节点来说，可以利用当前节 点的相邻节点的重建属性以及每个相邻节点距离当前节点的每个子节点的空间几何距离进行线性拟合，得到当前节点的每个子节点的预测属性；接着，利用每个子节点的预测属性进行RAHT属性变换得到相应的DC和AC系数，同时，可以通过RAHT属性变换来对当前节点的每个子节点的属性进行变换得到DC和AC系数；接着，可以利用预测节点得到的AC系数的预测值来对当前节点的AC系数进行预测，进而获得每个子节点的AC预测残差系数(系数残差)，然后可以对系数残差进行量化编码。另一方面，还可以利用AC预测残差系数的反量化残差值以及AC系数的预测值，恢复得到当前节点的AC重建系数(高频系数的重建值)，最终利用当前节点的AC系数和DC系数进行RAHT反变换，从而恢复得到当前节点每个子节点的属性重建值。
进一步地，在本申请实施例中，如果当前层为非跳过编码层，那么可以对当前层的节点依次进行变换、预测等过程，确定出当前层的节点对应的子节点的属性重建值；然后针对当前层节点的子节点来说，可以先确定子节点对应的下一层子节点的第六数量；然后根据第五数量和第六数量，确定子节点对应的下一层子节点的属性重建值。
也就是说，在本申请实施例中，无论当前层是否为跳过编码层，即无论是否对当前层的节点依次进行变换、预测等过程；仍然需要不断重复上述步骤，依然确定其它层的节点数量，以及对应的子节点数量，然后再根据节点数量和对应的子节点数量来判断是否执行跳过编码的处理。
相应地，在本申请实施例中，不断重复步骤S4601至步骤S4602的方法，依次从RAHT变换的根节点不断重复直至到RAHT的叶子结点层的最后一个节点，从而完成整个RAHT变换的属性编码。
进一步地，在一些实施例中，该方法还可以包括：确定预测模式标识信息；在预测模式标识信息指示当前单元启动跳过编码模式时，执行第一数量和第二数量的确定步骤，和/或，执行第四数量和第五数量的确定步骤。
在本申请实施例中，预测模式标识信息至少为以下高层语法元素中的其中之一：属性参数集(Attribute Parameter Set，APS)对应的语法元素和属性块头信息(Attribute Block Head，ABH)对应的语法元素。
在本申请实施例中，还可以对预测模式标识信息进行编码处理，将所得到的编码比特写入码流。其中，若当前单元启动跳过编码模式，则确定预测模式标识信息的取值为第一值；若当前单元不启动跳过编码模式，则确定预测模式标识信息的取值为第二值。下面针对当前单元的体素节点是否启动跳过编码模式和当前层的节点是否启动跳过编码模式分别进行说明。
在一种具体的实施例中，该方法还可以包括：确定第一预测模式标识信息；在第一预测模式标识信息指示当前单元的体素节点启动跳过编码模式时，执行第一数量和第二数量的确定步骤。
在本申请实施例中，还可以对第一预测模式标识信息进行编码处理，将所得到的编码比特写入码流。其中，若当前单元的体素节点启动跳过编码模式，则确定第一预测模式标识信息的取值为第一值；若当前单元的体素节点不启动跳过编码模式，则确定第一预测模式标识信息的取值为第二值。
还需要说明的是，在本申请实施例中，只有在当前单元的体素节点启动跳过编码模式时，这时候可以进一步确定第一数量和第二数量，然后根据第一数量和第二数量的大小来决定当前单元的体素节点是否跳过编码。具体地，如果两者一致，则跳过编码体素级别的节点属性，否则，会采用一个计时器，当该计时器的剩余重复节点数目为零时，则表示后续中的点不存在重复节点，同样可以跳过处理后续点的属性编码，并且将后续体素节点的属性重建值直接复制为最终剩余重建点的属性重建值。
在另一种具体的实施例中，该方法还可以包括：确定第二预测模式标识信息；在第二预测模式标识信息指示当前层的节点启动跳过编码模式时，执行第四数量和第五数量的确定步骤。
在本申请实施例中，还可以对第二预测模式标识信息进行编码处理，将所得到的编码比特写入码流。其中，若当前层的节点启动跳过编码模式，则确定第二预测模式标识信息的取值为第一值；若当前层的节点不启动跳过编码模式，则确定第二预测模式标识信息的取值为第二值。
还需要说明的是，在本申请实施例中，只有在当前层的节点启动跳过编码模式时，这时候可以进一步确定第四数量和第五数量，然后根据第四数量和第五数量的大小来决定当前层的节点是否跳过编码，即当前层是否为跳过编码层。具体地，如果两者一致，则确定当前层为跳过编码层，这时候不再需要编码当前层的节点对应的子节点的属性重建值；如果两者不一致，则确定当前层不属于跳过编码层，这时候可以按照相关技术的编码方式进行RAHT预测以及编码。
还需要说明的是，在本申请实施例中，第一值与第二值不同，而且第一值和第二值可以是参数形式，也可以是数字形式。具体地，第一预测模式标识信息和第二预测模式标识信息可以是写入在概述(profile)中的参数，也可以是一个标志(flag)的取值，这里不作具体限定。另外，对于第一值和第二值而言，第一值可以设置为1，第二值可以设置为0；或者，第一值可以设置为0，第二值可以设置为1；或者，第一值可以设置为true，第二值可以设置为false；或者，第一值可以设置为false，第二值可以设置为 true。其中，在本申请实施例中，第一值设置为1，第二值设置为0，但是不作具体限定。
示例性地，在一些实施例中，以第一值设置为1，第二值设置为0为例，如果当前层的节点启动跳过编码模式，那么即可确定第二预测模式标识信息的取值为1，进而可以根据上述方法进一步确定当前层的节点对应的第四数量和第五数量；如果当前层的节点不启动跳过编码模式，那么可以确定第二预测模式标识信息的取值为0，则可以按照常见的帧内预测的方式或帧间预测的方式对当前层的节点进行属性编码处理。
也就是说，在本申请实施例中，如果码流中写入的第一预测模式标识信息的取值为第一值，即确定当前单元的体素节点启动跳过编码模式，那么可以执行第一数量和第二数量的确定步骤，即执行图44所示的编码流程；如果码流中写入的第二预测模式标识信息的取值为第一值，即确定当前层的节点启动跳过编码模式，那么可以执行第四数量和第五数量的确定流程，即执行图46所示的编码流程。
进一步地，本申请实施例还提供了一种码流，码流是根据待编码信息进行比特编码生成的；其中，待编码信息可以包括下述至少一项：预测模式标识信息、当前单元的重复节点的属性信息和当前层的节点对应的量化后的系数残差；其中，预测模式标识信息用于指示当前单元是否启动跳过编码模式。
需要说明的是，在本申请实施例中，编码端在确定出待编码信息之后，可以将这些待编码信息写入码流；然后由编码端传输到解码端，后续在解码端，通过解码码流就可得到这些信息，例如预测模式标识信息，进而可以确定当前单元是否启动跳过编码模式。
综上所述，在本申请实施例中，在对属性信息进行编码时，如果当前层的节点数目与当前层的子节点数目一致，则认为当前层属于跳过编码层，因此不需要对当前层进行变换、预测、编码以及解码等过程，从而可以降低属性变换编解码的时间复杂度，并且对属性的编码效率不会产生任何影响。
本实施例提供了一种编码方法，这里提出了一种跳过当前层的编码方式，通过利用当前层的节点数目与子节点数目来自适应地决定当前层是否可以跳过编码，这样可以在保证编解码效率不变的前提下，降低编解码的时间复杂度；另外，这里还提出了一种跳过体素级别节点的编码方式，首先可以得到体素节点数目和重建节点数目，当重建节点数目与体素节点数目一致时，则认为当前单元中不存在重复节点，同样可以跳过编码体素级别的节点属性；如此，在保证点云属性的编解码效率基础上，可以降低点云属性编解码的时间复杂度，并且还可以节省码率，进而提升了点云的编解码性能。
在本申请的又一实施例中，基于前述实施例的编解码方法，本申请实施例首先定义RAHT属性编码层，目前的属性RAHT变换编码顺序是从根节点依次进行划分直至划分到体素级别(1×1×1)，从而完成整个点云的属性编码和属性重建。在这里，可以定义每次沿着Z方向、Y方向和X方向做一次下采样得到的层即为一个RAHT变换层，即layer。具体如图42所示。
其次，基于RAHT属性编码层，引入跳过编码层的算法。首先，在对当前层节点的属性进行编码/解码时，可以得到当前层的节点数目，以及利用当前层节点的几何信息进行恢复得到当前层节点的子节点之后，可以得到当前层节点的子节点数目，基于两者的大小来决定当前层是否属于跳过编码层：
如果当前层节点的数据与当前层节点子节点数目一致，则当前层属于跳过编码层；
否则，当前层为非跳过编码层。
这样，在得到跳过编码层的启动条件之后，编码端的一种具体算法如下：
步骤1：判断当前层的节点数目与当前层的子节点数目是否一致，如果一致，则当前层属于跳过编码层；否则，为非跳过编码层；
步骤2：如果当前层不属于跳过编码层，则按照相关技术的编码方式进行变换、预测以及编码。
步骤3：如果当前层属于跳过编码层，则直接跳过当前层至下一层。
步骤4：不断重复上述步骤，直至编码到体素级别。
步骤5：对于体素级别的节点，在对体素级别的节点进行属性编码之前，首先判断当前编码单元中体素级别的节点数目是否与重建节点的数目一致，如果一致，则跳过编码体素级别点的属性；否则，会采用一个计时器，当该计时器的剩余重复点数目为零时，则表示后续中的节点不存在重复点，同样可以跳过处理后续节点的属性编码。
在解码端，一种具体算法如下：
步骤1：判断当前层的节点数目与当前层的子节点数目是否一致，如果一致，则当前层属于跳过解码层；否则，为非跳过解码层；
步骤2：如果当前层不属于跳过解码层，则按照相关技术的解码方式进行预测以及解码。
步骤3：如果当前层属于跳过解码层，则直接跳过当前层至下一层。
步骤4：不断重复上述步骤，直至解码到体素级别。
步骤5：对于体素级别的节点，在对体素级别的节点进行属性解码之前，首先判断当前解码单元中体素级别的节点数目是否与重建节点的数目一致，如果一致，则跳过解码体素级别点的属性；否则，会采用一个计时器，当该计时器的剩余重复点数目为零时，则表示后续中的节点不存在重复点，同样可以跳过处理后续节点的属性解码，并且将后续剩余体素节点重建的属性重建值直接复制为最终剩余重建节点的属性重建值。
也就是说，在本申请实施例中，这里提出了一种跳过编码层的编码方式，通过利用当前层的节点数目与子节点数目来自适应地决定当前层是否可以跳过编码，这样可以在保证编解码效率不变的前提下，降低编解码的时间复杂度。同样的，在RAHT编码方案中，编解码端首先完成非节点层属性信息的编码以及解码(即节点的大小大于或等于1×1×1)，最终在对体素级别节点的属性信息进行编码和解码，原因是由于点云中会存在重复点，因此需要先完成非体素级别点的属性信息编码和解码，其次在完成体素级别点属性信息的编码和解码。如此，在编解码端，首先可以得到体素级别重建点的数目，当重建点的数目与所需要重建的节点数目一致时，则认为当前编码单元(例如slice)中不存在重复点，同样可以跳过编码/解码体素级别的点属性信息。
通过上述实施例，对前述实施例的具体实现进行详细阐述，从中可以看出，根据前述实施例的技术方案，本申请实施例在对属性信息进行RAHT编码时，在每一个RAHT属性编码层通过判断当前层的节点数目与当前层节点的子节点数目是否一致，来决定当前RAHT属性编码层是否属于跳过编码层。如果当前层的节点数目与当前层节点的子节点数目一致，则认为当前层属于跳过编码层，即不需要进行变换、预测、编码以及解码等过程，从而可以降低属性变换编码/解码的时间复杂度，并且对属性的编码效率不会产生任何影响。
在本申请的再一实施例中，基于前述实施例相同的发明构思，参见图47，其示出了本申请实施例提供的一种编码器的组成结构示意图。如图47所示，该编码器470可以包括：第一确定单元4701和编码单元4702，其中：
第一确定单元4701，配置为确定当前单元的体素节点的第一数量和当前单元的重建节点的第二数量；
编码单元4702，配置为根据第一数量与第二数量，确定当前单元的体素节点的属性信息是否跳过编码。
在一些实施例中，编码单元4702，还配置为若第一数量与第二数量相同，则跳过编码体素节点的属性信息；若第一数量与第二数量不同，则对体素节点中的重复节点进行属性编码，以及跳过编码体素节点中除重复节点之外剩余体素节点的属性信息。
在一些实施例中，编码单元4702，还配置为对重复节点的属性信息进行编码处理，将所得到的编码比特写入码流。
在一些实施例中，编码单元4702，还配置为设置计时器；在重复节点的属性信息开始编码时，启动计时器，并在计时器的计时达到预设数值时，确定重复节点的属性信息全部编码完成。
在一些实施例中，第一确定单元4701，还配置为根据第一数量与第二数量之间的差值，确定体素节点中的重复节点的第三数量；其中，计时器的设置与第三数量有关联关系。
在一些实施例中，参见图47，编码器470还可以包括第一重建单元4703，,配置为根据第一数量与第二数量，确定当前单元的体素节点的属性重建值。
在一些实施例中，第一重建单元4703，还配置为在第一数量与第二数量相同时，将当前单元的体素节点的属性重建值设置为当前单元的重建节点的属性重建值。
在一些实施例中，第一重建单元4703，还配置为在第一数量与第二数量不同时，将重复节点的属性重建值作为当前单元中第一重建节点的属性重建值，以及将体素节点中除重复节点之外的剩余体素节点的属性重建值设置为当前单元中除第一重建节点之外的剩余重建节点的属性重建值。
在一些实施例中，第一确定单元4701，还配置为对当前单元中的节点进行划分，确定至少一层；以及在最后一层的节点被划分到体素级别时，确定当前单元的体素节点以及体素节点的第一数量。
在一些实施例中，至少一层包括当前层，第一确定单元4701，还配置为确定当前层的节点的第四数量和当前层的节点对应的子节点的第五数量；其中，第四数量和第五数量用于确定是否对当前层进行跳过编码；
第一重建单元4703，还配置为根据第四数量和第五数量，确定当前层的节点对应的子节点的属性重建值。
在一些实施例中，第四数量表征当前层的被占据节点的数量；第五数量表征当前层的节点中被占据的子节点的数量或待编码的节点的数量。
在一些实施例中，第一确定单元4701，还配置为若第四数量和第五数量相同，则将当前层的节点的属性重建值确定为当前层的节点对应的子节点的属性重建值。
在一些实施例中，第一确定单元4701，还配置为若第四数量和第五数量相同，则确定子节点对应的下一层子节点的第六数量；
第一重建单元4703，还配置为根据第五数量和第六数量，确定子节点对应的下一层子节点的属性重建值。
在一些实施例中，第一重建单元4703，还配置为若第四数量和第五数量不同，则根据当前层的节点确定当前层的节点对应的子节点的属性预测值；基于子节点的属性预测值和子节点的属性值分别进行RAHT变换，确定当前层的节点对应的高频系数的重建值和低频系数；以及基于高频系数的重建值和低频系数进行RAHT逆变换，确定子节点的属性重建值。
在一些实施例中，第一确定单元4701，还配置为确定当前层的节点对应的相邻节点；以及根据相邻节点对应的属性重建值和相对距离参数，确定当前层的节点对应的子节点的属性预测值。
在一些实施例中，第一重建单元4703，还配置为基于子节点的属性预测值进行RAHT变换，确定当前层的节点对应的高频系数的预测值和低频系数；基于子节点的属性值进行RAHT变换，确定当前层的节点对应的高频系数和低频系数；以及根据当前层的节点对应的高频系数的预测值和当前层的节点对应的高频系数，确定当前层的节点对应的高频系数的重建值。
在一些实施例中，第一重建单元4703，还配置为根据当前层的节点对应的高频系数的预测值和当前层的节点对应的高频系数，确定当前层的节点对应的系数残差；以及根据当前层的节点对应的系数残差，确定当前层的节点对应的高频系数的重建值。
在一些实施例中，第一确定单元4701，还配置为对系数残差进行量化，确定当前层的节点对应的量化后的系数残差；
编码单元4702，还配置为对量化后的系数残差进行编码处理，将所得到的编码比特写入码流。
在一些实施例中，第一重建单元4703，还配置为对量化后的系数残差进行反量化，确定当前层的节点对应的反量化残差值；以及根据当前层的节点对应的反量化残差值和当前层的节点对应的高频系数的预测值，确定当前层的节点对应的高频系数的重建值。
在一些实施例中，第一确定单元4701，还配置为确定当前层的节点的几何信息；根据几何信息确定当前层的节点对应的子节点，以及第五数量。
在一些实施例中，第一确定单元4701，还配置为确定预测模式标识信息；
编码单元4702，还配置为在预测模式标识信息指示当前单元启动跳过编码模式时，执行第一数量和第二数量的确定步骤，和/或，执行第四数量和第五数量的确定步骤。
在一些实施例中，预测模式标识信息至少为以下高层语法元素中的其中之一：属性参数集对应的语法元素和属性块头信息对应的语法元素。
在一些实施例中，编码单元4702，还配置为对预测模式标识信息进行编码处理，将所得到的编码比特写入码流。
可以理解地，在本申请实施例中，“单元”可以是部分电路、部分处理器、部分程序或软件等等，当然也可以是模块，还可以是非模块化的。而且在本实施例中的各组成部分可以集成在一个处理单元中，也可以是各个单元单独物理存在，也可以两个或两个以上单元集成在一个单元中。上述集成的单元既可以采用硬件的形式实现，也可以采用软件功能模块的形式实现。
所述集成的单元如果以软件功能模块的形式实现并非作为独立的产品进行销售或使用时，可以存储在一个计算机可读取存储介质中，基于这样的理解，本实施例的技术方案本质上或者说对现有技术做出贡献的部分或者该技术方案的全部或部分可以以软件产品的形式体现出来，该计算机软件产品存储在一个存储介质中，包括若干指令用以使得一台计算机设备(可以是个人计算机，服务器，或者网络设备等)或processor(处理器)执行本实施例所述方法的全部或部分步骤。而前述的存储介质包括：U盘、移动硬盘、只读存储器(Read Only Memory，ROM)、随机存取存储器(Random Access Memory，RAM)、磁碟或者光盘等各种可以存储程序代码的介质。
因此，本申请实施例提供了一种计算机可读存储介质，应用于编码器470，该计算机可读存储介质存储有计算机程序，所述计算机程序被第一处理器执行时实现前述实施例中任一项所述的方法。
基于编码器470的组成以及计算机可读存储介质，参见图48，其示出了本申请实施例提供的编码器470的具体硬件结构示意图。如图48所示，编码器470可以包括：第一通信接口4801、第一存储器4802和第一处理器4803；各个组件通过第一总线系统4804耦合在一起。可理解，第一总线系统4804用于实现这些组件之间的连接通信。第一总线系统4804除包括数据总线之外，还包括电源总线、控制总线和状态信号总线。但是为了清楚说明起见，在图48中将各种总线都标为第一总线系统4804。其中：
第一通信接口4801，用于在与其他外部网元之间进行收发信息过程中，信号的接收和发送；
第一存储器4802，用于存储能够在第一处理器4803上运行的计算机程序；
第一处理器4803，用于在运行所述计算机程序时，执行：
确定当前单元的体素节点的第一数量和当前单元的重建节点的第二数量；
根据第一数量与第二数量，确定当前单元的体素节点的属性信息是否跳过编码。
可以理解，本申请实施例中的第一存储器4802可以是易失性存储器或非易失性存储器，或可包括易失性和非易失性存储器两者。其中，非易失性存储器可以是只读存储器(Read-Only Memory，ROM)、可编程只读存储器(Programmable ROM，PROM)、可擦除可编程只读存储器(Erasable PROM，EPROM)、电可擦除可编程只读存储器(Electrically EPROM，EEPROM)或闪存。易失性存储器可以是随机存取存储器(Random Access Memory，RAM)，其用作外部高速缓存。通过示例性但不是限制性说明，许多形式的RAM可用，例如静态随机存取存储器(Static RAM，SRAM)、动态随机存取存储器(Dynamic RAM，DRAM)、同步动态随机存取存储器(Synchronous DRAM，SDRAM)、双倍数据速率同步动态随机存取存储器(Double Data Rate SDRAM，DDRSDRAM)、增强型同步动态随机存取存储器(Enhanced SDRAM，ESDRAM)、同步连接动态随机存取存储器(Synchlink DRAM，SLDRAM)和直接内存总线随机存取存储器(Direct Rambus RAM，DRRAM)。本申请描述的系统和方法的第一存储器4802旨在包括但不限于这些和任意其它适合类型的存储器。
而第一处理器4803可能是一种集成电路芯片，具有信号的处理能力。在实现过程中，上述方法的各步骤可以通过第一处理器4803中的硬件的集成逻辑电路或者软件形式的指令完成。上述的第一处理器4803可以是通用处理器、数字信号处理器(Digital Signal Processor，DSP)、专用集成电路(Application Specific Integrated Circuit，ASIC)、现成可编程门阵列(Field Programmable Gate Array，FPGA)或者其他可编程逻辑器件、分立门或者晶体管逻辑器件、分立硬件组件。可以实现或者执行本申请实施例中的公开的各方法、步骤及逻辑框图。通用处理器可以是微处理器或者该处理器也可以是任何常规的处理器等。结合本申请实施例所公开的方法的步骤可以直接体现为硬件译码处理器执行完成，或者用译码处理器中的硬件及软件模块组合执行完成。软件模块可以位于随机存储器，闪存、只读存储器，可编程只读存储器或者电可擦写可编程存储器、寄存器等本领域成熟的存储介质中。该存储介质位于第一存储器4802，第一处理器4803读取第一存储器4802中的信息，结合其硬件完成上述方法的步骤。
可以理解的是，本申请描述的这些实施例可以用硬件、软件、固件、中间件、微码或其组合来实现。对于硬件实现，处理单元可以实现在一个或多个专用集成电路(Application Specific Integrated Circuits，ASIC)、数字信号处理器(Digital Signal Processing，DSP)、数字信号处理设备(DSP Device，DSPD)、可编程逻辑设备(Programmable Logic Device，PLD)、现场可编程门阵列(Field-Programmable Gate Array，FPGA)、通用处理器、控制器、微控制器、微处理器、用于执行本申请所述功能的其它电子单元或其组合中。对于软件实现，可通过执行本申请所述功能的模块(例如过程、函数等)来实现本申请所述的技术。软件代码可存储在存储器中并通过处理器执行。存储器可以在处理器中或在处理器外部实现。
可选地，作为另一个实施例，第一处理器4803还配置为在运行所述计算机程序时，执行前述实施例中任一项所述的方法。
本实施例提供了一种编码器，在该编码器中，在对每个体素节点进行属性重建时，优化了每个体素节点是否进行属性编解码的判断条件，具体为如果当前单元不存在重复节点，即第一数量与第二数量相同，此时不需要编解码当前单元的体素节点，从而在保证点云属性的编解码效率基础上，可以降低点云属性编解码的时间复杂度，并且还可以节省码率，进而提升了点云的编解码性能。
在本申请的再一实施例中，基于前述实施例相同的发明构思，参见图49，其示出了本申请实施例提供的一种解码器的组成结构示意图。如图49所示，该解码器490可以包括第二确定单元4901和第二重建单元4902，其中：
第二确定单元4901，配置为确定当前单元的体素节点的第一数量和当前单元的重建节点的第二数量；其中，第一数量与第二数量用于确定是否对当前单元的体素节点进行跳过解码；
第二重建单元4902，配置为根据第一数量与第二数量，确定当前单元的体素节点的属性重建值。
在一些实施例中，参见图49，解码器490还可以包括解码单元4903，配置为若第一数量与第二数量相同，则跳过解码当前单元的体素节点。
在一些实施例中，第二重建单元4902，还配置为在第一数量与第二数量相同时，将当前单元的体素节点的属性重建值设置为当前单元的重建节点的属性重建值。
在一些实施例中，解码单元4903，还配置为若第一数量与第二数量不同，则对体素节点中的重复节点进行属性解码，以及跳过解码体素节点中除重复节点之外的剩余体素节点。
在一些实施例中，第二重建单元4902，还配置为在第一数量与第二数量不同时，解码码流，确定重复节点的属性重建值；将重复节点的属性重建值作为当前单元中第一重建节点的属性重建值，以及将体素节点中除重复节点之外的剩余体素节点的属性重建值设置为当前单元中除第一重建节点之外的剩余重建节点的属性重建值。
在一些实施例中，解码单元4903，还配置为设置计时器；在重复节点的属性重建值开始解码时，启动计时器，并在计时器的计时达到预设数值时，确定重复节点的属性重建值全部解码完成。
在一些实施例中，在第一数量与第二数量不同时，第二确定单元4901，还配置为根据第一数量与第二数量之间的差值，确定体素节点中的重复节点的第三数量；其中，计时器的设置与第三数量有关联关系。
在一些实施例中，第二确定单元4901，还配置为对当前单元中的节点进行划分，确定至少一层；以及在最后一层的节点被划分到体素级别时，确定当前单元的体素节点以及体素节点的第一数量。
在一些实施例中，至少一层包括当前层，第二确定单元4901，还配置为确定当前层的节点的第四数量和当前层的节点对应的子节点的第五数量；其中，第四数量和第五数量用于确定是否对当前层进行跳过解码；
第二重建单元4902，还配置为根据第四数量和第五数量，确定当前层的节点对应的子节点的属性重建值。
在一些实施例中，第四数量表征当前层的被占据节点的数量；第五数量表征当前层的节点中被占据的子节点的数量或待解码的节点的数量。
在一些实施例中，第二重建单元4902，还配置为若第四数量和第五数量相同，则将当前层的节点的属性重建值确定为当前层的节点对应的子节点的属性重建值。
在一些实施例中，第二重建单元4902，还配置为若第四数量和第五数量相同，则确定子节点对应的下一层子节点的第六数量；根据第五数量和第六数量，确定子节点对应的下一层子节点的属性重建值。
在一些实施例中，第二重建单元4902，还配置为若第四数量和第五数量不同，则根据当前层的节点确定当前层的节点对应的子节点的属性预测值；基于子节点的属性预测值进行RAHT变换，确定当前层的节点对应的高频系数的重建值和低频系数；以及基于高频系数的重建值和低频系数进行RAHT逆变换，确定子节点的属性重建值。
在一些实施例中，第二确定单元4901，还配置为确定当前层的节点对应的相邻节点；以及根据相邻节点对应的属性重建值和相对距离参数，确定当前层的节点对应的子节点的属性预测值。
在一些实施例中，第二重建单元4902，还配置为基于子节点的属性预测值进行RAHT变换，确定当前层的节点对应的高频系数的预测值和低频系数；以及根据高频系数的预测值确定当前层的节点对应的高频系数的重建值。
在一些实施例中，解码单元4903，还配置为解码码流，确定当前层的节点对应的量化后的系数残差；
第二确定单元4901，还配置为根据高频系数的预测值和量化后的系数残差，确定当前层的节点对应的高频系数的重建值。
在一些实施例中，第二重建单元4902，还配置为对量化后的系数残差进行反量化，确定当前层的节点对应的反量化残差值；以及根据当前层的节点对应的反量化残差值和当前层的节点对应的高频系数的预测值，确定当前层的节点对应的高频系数的重建值。
在一些实施例中，第二确定单元4901，还配置为确定当前层的节点的几何信息；以及根据几何信息确定当前层的节点对应的子节点，以及第五数量。
在一些实施例中，解码单元4903，还配置为解码码流，确定预测模式标识信息；以及在预测模式标识信息指示当前单元启动跳过解码模式时，执行第一数量和第二数量的确定步骤，和/或，执行第四数量和第五数量的确定步骤。
在一些实施例中，预测模式标识信息至少为以下高层语法元素中的其中之一：属性参数集对应的语法元素和属性块头信息对应的语法元素。
可以理解地，在本实施例中，“单元”可以是部分电路、部分处理器、部分程序或软件等等，当然也可以是模块，还可以是非模块化的。而且在本实施例中的各组成部分可以集成在一个处理单元中，也可以是各个单元单独物理存在，也可以两个或两个以上单元集成在一个单元中。上述集成的单元既可以采用硬件的形式实现，也可以采用软件功能模块的形式实现。
所述集成的单元如果以软件功能模块的形式实现并非作为独立的产品进行销售或使用时，可以存储在一个计算机可读取存储介质中。基于这样的理解，本实施例提供了一种计算机可读存储介质，应用于解码器490，该计算机可读存储介质存储有计算机程序，所述计算机程序被第二处理器执行时实现前述 实施例中任一项所述的方法。
基于解码器490的组成以及计算机可读存储介质，参见图50，其示出了本申请实施例提供的解码器490的具体硬件结构示意图。如图50所示，解码器490可以包括：第二通信接口5001、第二存储器5002和第二处理器5003；各个组件通过第二总线系统5004耦合在一起。可理解，第二总线系统5004用于实现这些组件之间的连接通信。第二总线系统5004除包括数据总线之外，还包括电源总线、控制总线和状态信号总线。但是为了清楚说明起见，在图50中将各种总线都标为第二总线系统5004。其中：
第二通信接口5001，用于在与其他外部网元之间进行收发信息过程中，信号的接收和发送；
第二存储器5002，用于存储能够在第二处理器5003上运行的计算机程序；
第二处理器5003，用于在运行所述计算机程序时，执行：
确定当前单元的体素节点的第一数量和当前单元的重建节点的第二数量；其中，第一数量与第二数量用于确定是否对当前单元的体素节点进行跳过解码；
根据第一数量与第二数量，确定当前单元的体素节点的属性重建值。
可选地，作为另一个实施例，第二处理器5003还配置为在运行所述计算机程序时，执行前述实施例中任一项所述的方法。
可以理解，第二存储器5002与第一存储器4802的硬件功能类似，第二处理器5003与第一处理器4803的硬件功能类似；这里不再详述。
本实施例提供了一种解码器，在该解码器中，在对每个体素节点进行属性重建时，优化了每个体素节点是否进行属性编解码的判断条件，具体为如果当前单元不存在重复节点，即第一数量与第二数量相同，此时不需要编解码当前单元的体素节点，从而在保证点云属性的编解码效率基础上，可以降低点云属性编解码的时间复杂度，并且还可以节省码率，进而提升了点云的编解码性能。
在本申请的再一实施例中，参见图51，其示出了本申请实施例提供的一种编解码系统的组成结构示意图。如图51所示，编解码系统510可以包括编码器5101和解码器5102。
在本申请实施例中，编码器5101可以是前述实施例中任一项所述的编码器，解码器5102可以是前述实施例中任一项所述的解码器。
需要说明的是，在本申请中，术语“包括”、“包含”或者其任何其他变体意在涵盖非排他性的包含，从而使得包括一系列要素的过程、方法、物品或者装置不仅包括那些要素，而且还包括没有明确列出的其他要素，或者是还包括为这种过程、方法、物品或者装置所固有的要素。在没有更多限制的情况下，由语句“包括一个……”限定的要素，并不排除在包括该要素的过程、方法、物品或者装置中还存在另外的相同要素。
上述本申请实施例序号仅仅为了描述，不代表实施例的优劣。
本申请所提供的几个方法实施例中所揭露的方法，在不冲突的情况下可以任意组合，得到新的方法实施例。
本申请所提供的几个产品实施例中所揭露的特征，在不冲突的情况下可以任意组合，得到新的产品实施例。
本申请所提供的几个方法或设备实施例中所揭露的特征，在不冲突的情况下可以任意组合，得到新的方法实施例或设备实施例。
以上所述，仅为本申请的具体实施方式，但本申请的保护范围并不局限于此，任何熟悉本技术领域的技术人员在本申请揭露的技术范围内，可轻易想到变化或替换，都应涵盖在本申请的保护范围之内。因此，本申请的保护范围应以所述权利要求的保护范围为准。
工业实用性
本申请实施例中，无论是编码端还是解码端，确定当前单元的体素节点的第一数量和当前单元的重建节点的第二数量，然后根据第一数量与第二数量，确定当前单元的体素节点的属性重建值；其中，第一数量与第二数量用于确定当前单元的体素节点是否跳过编解码。以及确定当前层的节点的第四数量和当前层的节点对应的子节点的第五数量；根据第四数量和第五数量，确定当前层的节点对应的子节点的属性重建值；其中，第四数量和第五数量用于确定当前层是否跳过编解码。这样，可以利用当前层的节点数目与子节点数目来自适应地决定当前层是否可以跳过解码，还可以在重建节点数目与体素节点数目一致时，认为当前单元中不存在重复节点，同样可以跳过解码体素级别的节点属性；如此，在保证点云属性的编解码效率基础上，可以降低点云属性编解码的时间复杂度，并且还可以节省码率，进而提升了点云的编解码性能。

Claims (49)

  1. 一种解码方法，应用于解码器，所述方法包括：
    确定当前单元的体素节点的第一数量和所述当前单元的重建节点的第二数量；其中，所述第一数量与所述第二数量用于确定是否对所述当前单元的体素节点进行跳过解码；
    根据所述第一数量与所述第二数量，确定所述当前单元的体素节点的属性重建值。
  2. 根据权利要求1所述的方法，其中，所述方法还包括：
    若所述第一数量与所述第二数量相同，则跳过解码所述当前单元的体素节点。
  3. 根据权利要求2所述的方法，其中，所述根据所述第一数量与所述第二数量，确定所述当前单元的体素节点的属性重建值，包括：
    在所述第一数量与所述第二数量相同时，将所述当前单元的体素节点的属性重建值设置为所述当前单元的重建节点的属性重建值。
  4. 根据权利要求1所述的方法，其中，所述方法还包括：
    若所述第一数量与所述第二数量不同，则对所述体素节点中的重复节点进行属性解码，以及跳过解码所述体素节点中除所述重复节点之外的剩余体素节点。
  5. 根据权利要求4所述的方法，其中，所述根据所述第一数量与所述第二数量，确定所述当前单元的体素节点的属性重建值，包括：
    在所述第一数量与所述第二数量不同时，解码码流，确定所述重复节点的属性重建值；
    将所述重复节点的属性重建值作为所述当前单元中第一重建节点的属性重建值，以及将所述体素节点中除所述重复节点之外的剩余体素节点的属性重建值设置为所述当前单元中除所述第一重建节点之外的剩余重建节点的属性重建值。
  6. 根据权利要求5所述的方法，其中，所述解码码流，确定所述重复节点的属性重建值，包括：
    设置计时器；
    在所述重复节点的属性重建值开始解码时，启动所述计时器，并在所述计时器的计时达到预设数值时，确定所述重复节点的属性重建值全部解码完成。
  7. 根据权利要求6所述的方法，其中，在所述第一数量与所述第二数量不同时，所述方法还包括：
    根据所述第一数量与所述第二数量之间的差值，确定所述体素节点中的重复节点的第三数量；其中，所述计时器的设置与所述第三数量有关联关系。
  8. 根据权利要求1至7中任一项所述的方法，其中，所述确定当前单元的体素节点的第一数量，包括：
    对所述当前单元中的节点进行划分，确定至少一层；
    在最后一层的节点被划分到体素级别时，确定所述当前单元的体素节点以及所述体素节点的第一数量。
  9. 根据权利要求8所述的方法，其中，所述至少一层包括当前层，所述方法还包括：
    确定所述当前层的节点的第四数量和所述当前层的节点对应的子节点的第五数量；其中，所述第四数量和所述第五数量用于确定是否对所述当前层进行跳过解码；
    根据所述第四数量和所述第五数量，确定所述当前层的节点对应的子节点的属性重建值。
  10. 根据权利要求9所述的方法，其中，
    所述第四数量表征所述当前层的被占据节点的数量；
    所述第五数量表征所述当前层的节点中被占据的子节点的数量或待解码的节点的数量。
  11. 根据权利要求9所述的方法，其中，所述根据所述第四数量和所述第五数量，确定所述当前层的节点对应的子节点的属性重建值，包括：
    若所述第四数量和所述第五数量相同，则将所述当前层的节点的属性重建值确定为所述当前层的节点对应的子节点的属性重建值。
  12. 根据权利要求9所述的方法，其中，所述根据所述第四数量和所述第五数量，确定所述当前层的节点对应的子节点的属性重建值，包括：
    若所述第四数量和所述第五数量相同，则确定所述子节点对应的下一层子节点的第六数量；
    根据所述第五数量和所述第六数量，确定所述子节点对应的下一层子节点的属性重建值。
  13. 根据权利要求9所述的方法，其中，所述根据所述第四数量和所述第五数量，确定所述当前层的节点对应的子节点的属性重建值，包括：
    若所述第四数量和所述第五数量不同，则根据所述当前层的节点确定所述当前层的节点对应的子节 点的属性预测值；
    基于所述子节点的属性预测值进行RAHT变换，确定所述当前层的节点对应的高频系数的重建值和低频系数；
    基于所述高频系数的重建值和所述低频系数进行RAHT逆变换，确定所述子节点的属性重建值。
  14. 根据权利要求13所述的方法，其中，所述根据所述当前层的节点确定所述当前层的节点对应的子节点的属性预测值，包括：
    确定所述当前层的节点对应的相邻节点；
    根据所述相邻节点对应的属性重建值和相对距离参数，确定所述当前层的节点对应的子节点的属性预测值。
  15. 根据权利要求13所述的方法，其中，所述基于所述子节点的属性预测值进行RAHT变换，确定所述当前层的节点对应的高频系数的重建值和低频系数，包括：
    基于所述子节点的属性预测值进行RAHT变换，确定所述当前层的节点对应的高频系数的预测值和所述低频系数；
    根据所述高频系数的预测值确定所述当前层的节点对应的所述高频系数的重建值。
  16. 根据权利要求15所述的方法，其中，所述根据所述高频系数的预测值确定所述当前层的节点对应的所述高频系数的重建值，包括：
    解码码流，确定所述当前层的节点对应的量化后的系数残差；
    根据所述高频系数的预测值和所述量化后的系数残差，确定所述当前层的节点对应的所述高频系数的重建值。
  17. 根据权利要求16所述的方法，其中，所述根据所述高频系数的预测值和所述量化后系数残差，确定所述当前层的节点对应的所述高频系数的重建值，包括：
    对所述量化后的系数残差进行反量化，确定所述当前层的节点对应的反量化残差值；
    根据所述当前层的节点对应的反量化残差值和所述当前层的节点对应的高频系数的预测值，确定所述当前层的节点对应的高频系数的重建值。
  18. 根据权利要求9所述的方法，其中，所述方法还包括：
    确定所述当前层的节点的几何信息；
    根据所述几何信息确定所述当前层的节点对应的子节点，以及所述第五数量。
  19. 根据权利要求9所述的方法，其中，所述方法还包括：
    解码码流，确定预测模式标识信息；
    在所述预测模式标识信息指示所述当前单元启动跳过解码模式时，执行所述第一数量和所述第二数量的确定步骤，和/或，执行所述第四数量和所述第五数量的确定步骤。
  20. 根据权利要求19所述的方法，其中，所述预测模式标识信息至少为以下高层语法元素中的其中之一：属性参数集对应的语法元素和属性块头信息对应的语法元素。
  21. 一种编码方法，应用于编码器，所述方法包括：
    确定当前单元的体素节点的第一数量和所述当前单元的重建节点的第二数量；
    根据所述第一数量与所述第二数量，确定所述当前单元的体素节点的属性信息是否跳过编码。
  22. 根据权利要求21所述的方法，其中，所述根据所述第一数量与所述第二数量，确定所述当前单元的体素节点的属性信息是否跳过编码，包括：
    若所述第一数量与所述第二数量相同，则跳过编码所述体素节点的属性信息；
    若所述第一数量与所述第二数量不同，则对所述体素节点中的重复节点进行属性编码，以及跳过编码所述体素节点中除所述重复节点之外剩余体素节点的属性信息。
  23. 根据权利要求22所述的方法，其中，所述对所述体素节点中的重复节点进行属性编码，包括：
    对所述重复节点的属性信息进行编码处理，将所得到的编码比特写入码流。
  24. 根据权利要求23所述的方法，其中，所述对所述重复节点的属性信息进行编码处理，包括：
    设置计时器；
    在所述重复节点的属性信息开始编码时，启动所述计时器，并在所述计时器的计时达到预设数值时，确定所述重复节点的属性信息全部编码完成。
  25. 根据权利要求24所述的方法，其中，所述方法还包括：
    根据所述第一数量与所述第二数量之间的差值，确定所述体素节点中的重复节点的第三数量；其中，所述计时器的设置与所述第三数量有关联关系。
  26. 根据权利要求21所述的方法，其中，所述方法还包括：
    根据所述第一数量与所述第二数量，确定所述当前单元的体素节点的属性重建值。
  27. 根据权利要求26所述的方法，其中，所述根据所述第一数量与所述第二数量，确定所述当前单元的体素节点的属性重建值，包括：
    在所述第一数量与所述第二数量相同时，将所述当前单元的体素节点的属性重建值设置为所述当前单元的重建节点的属性重建值。
  28. 根据权利要求26所述的方法，其中，所述根据所述第一数量与所述第二数量，确定所述当前单元的体素节点的属性重建值，包括：
    在所述第一数量与所述第二数量不同时，将所述重复节点的属性重建值作为所述当前单元中第一重建节点的属性重建值，以及将所述体素节点中除所述重复节点之外的剩余体素节点的属性重建值设置为所述当前单元中除所述第一重建节点之外的剩余重建节点的属性重建值。
  29. 根据权利要求21至28中任一项所述的方法，其中，所述确定当前单元的体素节点的第一数量，包括：
    对所述当前单元中的节点进行划分，确定至少一层；
    在最后一层的节点被划分到体素级别时，确定所述当前单元的体素节点以及所述体素节点的第一数量。
  30. 根据权利要求29所述的方法，其中，所述至少一层包括当前层，所述方法还包括：
    确定所述当前层的节点的第四数量和所述当前层的节点对应的子节点的第五数量；其中，所述第四数量和所述第五数量用于确定是否对所述当前层进行跳过编码；
    根据所述第四数量和所述第五数量，确定所述当前层的节点对应的子节点的属性重建值。
  31. 根据权利要求30所述的方法，其中，
    所述第四数量表征所述当前层的被占据节点的数量；
    所述第五数量表征所述当前层的节点中被占据的子节点的数量或待编码的节点的数量。
  32. 根据权利要求30所述的方法，其中，所述根据所述第四数量和所述第五数量，确定所述当前层的节点对应的子节点的属性重建值，包括：
    若所述第四数量和所述第五数量相同，则将所述当前层的节点的属性重建值确定为所述当前层的节点对应的子节点的属性重建值。
  33. 根据权利要求30所述的方法，其中，所述根据所述第四数量和所述第五数量，确定所述当前层的节点对应的子节点的属性重建值，包括：
    若所述第四数量和所述第五数量相同，则确定所述子节点对应的下一层子节点的第六数量；
    根据所述第五数量和所述第六数量，确定所述子节点对应的下一层子节点的属性重建值。
  34. 根据权利要求30所述的方法，其中，所述根据所述第四数量和所述第五数量，确定所述当前层的节点对应的子节点的属性重建值，包括：
    若所述第四数量和所述第五数量不同，则根据所述当前层的节点确定所述当前层的节点对应的子节点的属性预测值；
    基于所述子节点的属性预测值和所述子节点的属性值分别进行RAHT变换，确定所述当前层的节点对应的高频系数的重建值和低频系数；
    基于所述高频系数的重建值和所述低频系数进行RAHT逆变换，确定所述子节点的属性重建值。
  35. 根据权利要求34所述的方法，其中，所述根据所述当前层的节点确定所述当前层的节点对应的子节点的属性预测值，包括：
    确定所述当前层的节点对应的相邻节点；
    根据所述相邻节点对应的属性重建值和相对距离参数，确定所述当前层的节点对应的子节点的属性预测值。
  36. 根据权利要求34所述的方法，其中，所述基于所述子节点的属性预测值和所述子节点的属性值分别进行RAHT变换，确定所述当前层的节点对应的高频系数的重建值和低频系数，包括：
    基于所述子节点的属性预测值进行RAHT变换，确定所述当前层的节点对应的高频系数的预测值和所述低频系数；
    基于所述子节点的属性值进行RAHT变换，确定所述当前层的节点对应的高频系数和所述低频系数；
    根据所述当前层的节点对应的高频系数的预测值和所述当前层的节点对应的高频系数，确定所述当前层的节点对应的高频系数的重建值。
  37. 根据权利要求36所述的方法，其中，所述根据所述当前层的节点对应的高频系数的预测值和所述当前层的节点对应的高频系数，确定所述当前层的节点对应的高频系数的重建值，包括：
    根据所述当前层的节点对应的高频系数的预测值和所述当前层的节点对应的高频系数，确定所述当 前层的节点对应的系数残差；
    根据所述当前层的节点对应的系数残差，确定所述当前层的节点对应的高频系数的重建值。
  38. 根据权利要求37所述的方法，其中，所述方法还包括：
    对所述系数残差进行量化，确定所述当前层的节点对应的量化后的系数残差；
    对所述量化后的系数残差进行编码处理，将所得到的编码比特写入码流。
  39. 根据权利要求38所述的方法，其中，所述根据所述当前层的节点对应的系数残差，确定所述当前层的节点对应的高频系数的重建值，包括：
    对所述量化后的系数残差进行反量化，确定所述当前层的节点对应的反量化残差值；
    根据所述当前层的节点对应的反量化残差值和所述当前层的节点对应的高频系数的预测值，确定所述当前层的节点对应的高频系数的重建值。
  40. 根据权利要求30所述的方法，其中，所述方法还包括：
    确定所述当前层的节点的几何信息；
    根据所述几何信息确定所述当前层的节点对应的子节点，以及所述第五数量。
  41. 根据权利要求30所述的方法，其中，所述方法还包括：
    确定预测模式标识信息；
    在所述预测模式标识信息指示所述当前单元启动跳过编码模式时，执行所述第一数量和所述第二数量的确定步骤，和/或，执行所述第四数量和所述第五数量的确定步骤。
  42. 根据权利要求41所述的方法，其中，所述预测模式标识信息至少为以下高层语法元素中的其中之一：属性参数集对应的语法元素和属性块头信息对应的语法元素。
  43. 根据权利要求41所述的方法，其中，所述方法还包括：
    对所述预测模式标识信息进行编码处理，将所得到的编码比特写入码流。
  44. 一种码流，其中，所述码流是根据待编码信息进行比特编码生成的；其中，待编码信息包括下述至少一项：
    预测模式标识信息、当前单元的重复节点的属性信息和当前层的节点对应的量化后的系数残差；其中，所述预测模式标识信息用于指示所述当前单元是否启动跳过编码模式。
  45. 一种编码器，所述编码器包括第一确定单元和编码单元，其中：
    所述第一确定单元，配置为确定当前单元的体素节点的第一数量和所述当前单元的重建节点的第二数量；
    所述编码单元，配置为根据所述第一数量与所述第二数量，确定所述当前单元的体素节点的属性信息是否跳过编码。
  46. 一种编码器，所述编码器包括第一存储器和第一处理器，其中：
    所述第一存储器，用于存储能够在所述第一处理器上运行的计算机程序；
    所述第一处理器，用于在运行所述计算机程序时，执行如权利要求21至43中任一项所述的方法。
  47. 一种解码器，所述解码器包括第二确定单元和第二重建单元，其中：
    所述第二确定单元，配置为确定当前单元的体素节点的第一数量和所述当前单元的重建节点的第二数量；其中，所述第一数量与所述第二数量用于确定是否对所述当前单元的体素节点进行跳过解码；
    所述第二重建单元，配置为根据所述第一数量与所述第二数量，确定所述当前单元的体素节点的属性重建值。
  48. 一种解码器，所述解码器包括第二存储器和第二处理器，其中：
    所述第二存储器，用于存储能够在所述第二处理器上运行的计算机程序；
    所述第二处理器，用于在运行所述计算机程序时，执行如权利要求1至20中任一项所述的方法。
  49. 一种计算机可读存储介质，其中，所述计算机可读存储介质存储有计算机程序，所述计算机程序被执行时实现如权利要求1至20中任一项所述的方法、或者实现如权利要求21至43中任一项所述的方法。
PCT/CN2023/106178 2023-07-06 2023-07-06 编解码方法、码流、编码器、解码器以及存储介质 Ceased WO2025007355A1 (zh)

Priority Applications (3)

Application Number Priority Date Filing Date Title
PCT/CN2023/106178 WO2025007355A1 (zh) 2023-07-06 2023-07-06 编解码方法、码流、编码器、解码器以及存储介质
CN202380098588.6A CN121241567A (zh) 2023-07-06 2023-07-06 编解码方法、码流、编码器、解码器以及存储介质
US19/423,914 US20260113485A1 (en) 2023-07-06 2025-12-17 Encoding method, decoding method, code stream, encoder, decoder, and storage medium

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2023/106178 WO2025007355A1 (zh) 2023-07-06 2023-07-06 编解码方法、码流、编码器、解码器以及存储介质

Related Child Applications (1)

Application Number Title Priority Date Filing Date
US19/423,914 Continuation US20260113485A1 (en) 2023-07-06 2025-12-17 Encoding method, decoding method, code stream, encoder, decoder, and storage medium

Publications (2)

Publication Number Publication Date
WO2025007355A1 true WO2025007355A1 (zh) 2025-01-09
WO2025007355A9 WO2025007355A9 (zh) 2025-02-13

Family

ID=94171050

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2023/106178 Ceased WO2025007355A1 (zh) 2023-07-06 2023-07-06 编解码方法、码流、编码器、解码器以及存储介质

Country Status (3)

Country Link
US (1) US20260113485A1 (zh)
CN (1) CN121241567A (zh)
WO (1) WO2025007355A1 (zh)

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2022126326A1 (zh) * 2020-12-14 2022-06-23 Oppo广东移动通信有限公司 点云编解码方法、编码器、解码器以及计算机存储介质
CN115023741A (zh) * 2020-01-09 2022-09-06 松下电器（美国）知识产权公司 三维数据编码方法、三维数据解码方法、三维数据编码装置、以及三维数据解码装置
WO2022217215A1 (en) * 2021-04-05 2022-10-13 Qualcomm Incorporated Residual coding for geometry point cloud compression
CN115380537A (zh) * 2020-04-13 2022-11-22 Lg电子株式会社 发送点云数据的设备、发送点云数据的方法、接收点云数据的设备及接收点云数据的方法
CN116033186A (zh) * 2022-12-30 2023-04-28 腾讯科技（深圳）有限公司 一种点云数据处理方法、装置、设备以及介质

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115023741A (zh) * 2020-01-09 2022-09-06 松下电器（美国）知识产权公司 三维数据编码方法、三维数据解码方法、三维数据编码装置、以及三维数据解码装置
CN115380537A (zh) * 2020-04-13 2022-11-22 Lg电子株式会社 发送点云数据的设备、发送点云数据的方法、接收点云数据的设备及接收点云数据的方法
WO2022126326A1 (zh) * 2020-12-14 2022-06-23 Oppo广东移动通信有限公司 点云编解码方法、编码器、解码器以及计算机存储介质
WO2022217215A1 (en) * 2021-04-05 2022-10-13 Qualcomm Incorporated Residual coding for geometry point cloud compression
CN116033186A (zh) * 2022-12-30 2023-04-28 腾讯科技（深圳）有限公司 一种点云数据处理方法、装置、设备以及介质

Also Published As

Publication number Publication date
US20260113485A1 (en) 2026-04-23
WO2025007355A9 (zh) 2025-02-13
CN121241567A (zh) 2025-12-30

Similar Documents

Publication Publication Date Title
WO2024145904A1 (zh) 编解码方法、码流、编码器、解码器以及存储介质
WO2024145910A1 (zh) 编解码方法、码流、编码器、解码器以及存储介质
WO2025007355A1 (zh) 编解码方法、码流、编码器、解码器以及存储介质
WO2024216476A1 (zh) 编解码方法、编码器、解码器、码流以及存储介质
WO2024216477A1 (zh) 编解码方法、编码器、解码器、码流以及存储介质
WO2025010600A9 (zh) 编解码方法、码流、编码器、解码器以及存储介质
WO2024216479A9 (zh) 编解码方法、码流、编码器、解码器以及存储介质
WO2025010601A9 (zh) 编解码方法、编码器、解码器、码流以及存储介质
WO2025010604A1 (zh) 点云编解码方法、编码器、解码器、码流以及存储介质
WO2025076668A9 (zh) 编解码方法、编码器、解码器以及存储介质
WO2024234132A9 (zh) 编解码方法、码流、编码器、解码器以及存储介质
WO2024207456A1 (zh) 编解码方法、编码器、解码器、码流以及存储介质
WO2025145433A1 (zh) 点云编解码方法、编解码器、码流以及存储介质
WO2025007349A1 (zh) 编解码方法、码流、编码器、解码器以及存储介质
WO2025007360A1 (zh) 编解码方法、码流、编码器、解码器以及存储介质
WO2024207481A1 (zh) 编解码方法、编码器、解码器、码流以及存储介质
WO2025076672A1 (zh) 编解码方法、编码器、解码器、码流以及存储介质
US20260129189A1 (en) Encoding method, decoding method, encoders, decoders, bitstream and storage medium
WO2025076663A1 (zh) 编解码方法、编解码器以及存储介质
WO2024212038A1 (zh) 编解码方法、码流、编码器、解码器以及存储介质
WO2025145330A1 (zh) 点云编解码方法、编解码器、码流以及存储介质
WO2024148598A1 (zh) 编解码方法、编码器、解码器以及存储介质
WO2024212045A1 (zh) 编解码方法、码流、编码器、解码器以及存储介质
WO2025015523A1 (zh) 编解码方法、码流、编码器、解码器以及存储介质
WO2024212043A1 (zh) 编解码方法、码流、编码器、解码器以及存储介质

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 23944082

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE