WO2011047579A1 - 一种视频编解码方法及设备 - Google Patents

一种视频编解码方法及设备 Download PDF

Info

Publication number
WO2011047579A1
WO2011047579A1 PCT/CN2010/076464 CN2010076464W WO2011047579A1 WO 2011047579 A1 WO2011047579 A1 WO 2011047579A1 CN 2010076464 W CN2010076464 W CN 2010076464W WO 2011047579 A1 WO2011047579 A1 WO 2011047579A1
Authority
WO
WIPO (PCT)
Prior art keywords
transform
transformation
index information
matrices
matrix
Prior art date
Application number
PCT/CN2010/076464
Other languages
English (en)
French (fr)
Inventor
杨名远
王栋
熊联欢
赵欣
张莉
马思伟
高文
Original Assignee
华为技术有限公司
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 华为技术有限公司 filed Critical 华为技术有限公司
Priority to KR1020127011936A priority Critical patent/KR101481642B1/ko
Priority to EP10824420.3A priority patent/EP2493197A4/en
Priority to BR112012011325-9A priority patent/BR112012011325B1/pt
Priority to AU2010310286A priority patent/AU2010310286B2/en
Publication of WO2011047579A1 publication Critical patent/WO2011047579A1/zh
Priority to US13/452,198 priority patent/US9723313B2/en

Links

Classifications

    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/10Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
    • H04N19/134Methods 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
    • H04N19/136Incoming video signal characteristics or properties
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/10Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
    • H04N19/102Methods 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
    • H04N19/103Selection of coding mode or of prediction mode
    • H04N19/11Selection of coding mode or of prediction mode among a plurality of spatial predictive coding modes
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/10Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
    • H04N19/102Methods 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
    • H04N19/12Selection from among a plurality of transforms or standards, e.g. selection between discrete cosine transform [DCT] and sub-band transform or selection between H.263 and H.264
    • H04N19/122Selection of transform size, e.g. 8x8 or 2x4x8 DCT; Selection of sub-band transforms of varying structure or type
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/10Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
    • H04N19/134Methods 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
    • H04N19/157Assigned coding mode, i.e. the coding mode being predefined or preselected to be further used for selection of another element or parameter
    • H04N19/159Prediction type, e.g. intra-frame, inter-frame or bidirectional frame prediction
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/10Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
    • H04N19/169Methods 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
    • H04N19/17Methods 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 an image region, e.g. an object
    • H04N19/176Methods 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 an image region, e.g. an object the region being a block, e.g. a macroblock
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/46Embedding additional information in the video signal during the compression process
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/60Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using transform coding
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/60Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using transform coding
    • H04N19/61Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using transform coding in combination with predictive coding
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/10Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
    • H04N19/134Methods 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
    • H04N19/146Data rate or code amount at the encoder output
    • H04N19/147Data rate or code amount at the encoder output according to rate distortion criteria

Definitions

  • a video encoding and decoding method and device The present application claims to be submitted on October 23, 2009, the application number is 200910209013. 9, the invention name is "a video encoding and decoding method and device", submitted on April 9, 2010. The application number is 201010147581. 3, the invention name is “a video codec method and device” and submitted on June 17, 2010, the application number is 201010213791. 8 , the invention name is "a video codec method and device The priority of the Chinese Patent Application, the entire contents of which is incorporated herein by reference.
  • the present invention relates to the field of communications, and in particular, to a video encoding and decoding method and device. Background technique
  • a complete video codec system consists of an encoder and a decoder.
  • the video signal first passes through a prediction module, and the encoder selects the best one from several prediction modes according to certain optimization criteria, and then generates a residual signal; After the signal is transformed and quantized, it enters the entropy coding module and finally forms an output code stream.
  • the prediction mode information is first parsed from the code stream to generate a prediction signal that is completely consistent with the encoding end; then the transformed coefficient values in the code stream are parsed, and inverse quantization and inverse transformation are performed to generate a reconstructed residual. The difference signal; finally, the reconstructed video signal is synthesized by using the prediction signal and the reconstructed residual signal.
  • the coding process contains a key technology: transformation.
  • the function of the transformation is to transform the residual into another expression by performing some linear operation on the residual block, and in this expression, the energy of the data is concentrated on a few transform coefficients, and most of the rest. The energy of the coefficient is very low or zero, and by such conversion, subsequent entropy coding can be performed efficiently.
  • F C ⁇ X ⁇ R
  • C And R is the transformation matrix of the same size as X
  • F is the transformation coefficient matrix obtained by the transformation. Since the Di s crete Cos ine Transform (DCT) has a better compromise between the complexity and the performance than other existing transforms, it is widely accepted.
  • DCT Di s crete Cos ine Transform
  • an item is called dependent direction transform mode.
  • the technology of Di rect iona l Transform, MDDT was adopted.
  • the core idea is: 1 Because the residuals obtained by different intra prediction modes embody different statistical characteristics, the transform should use different transform matrices to improve the compression coding efficiency according to the different prediction directions. 2
  • i is the corresponding intra prediction mode.
  • J is the prediction residual
  • W is the predicted residual after the transformation, 67 and ⁇ 7 can be seen, the horizontal and vertical transformations are separated from the two matrices, which is the so-called transformation of the row and column separation.
  • the inventors have found that at least the following problems exist in the prior art: Although the MDDT technology can use different sets of transformation matrices for different prediction directions for intra coding, in the actual coding process. Even in a same intra prediction mode, the statistical characteristics of the residual data still have significant differences, so the above method of intra prediction mode corresponding to a set of transformation matrices is still not accurate enough, so that subsequent coding efficiency is low. . Summary of the invention
  • the embodiments of the present invention provide a video encoding and decoding method and device, which can perform a transform by selecting an effective transform matrix according to the characteristics of each residual block, thereby improving coding efficiency.
  • a video data encoding method includes:
  • a set of optimal transform matrices are selected from a plurality of candidate transform matrices according to a rate distortion criterion to transform and encode the prediction residuals to obtain a transform result;
  • the encoded code stream is generated based on the transformed result and the selected transform matrix index information.
  • a video data encoder comprising:
  • a residual generating unit configured to generate a prediction residual according to the input video data
  • a transforming unit configured to perform transform coding on the prediction residual by selecting a set of optimal transform matrices from the plurality of candidate transform matrices according to a rate prediction criterion according to a rate prediction criterion, to obtain a transform result;
  • a video data decoding method includes:
  • the calculation result is inversely transformed to obtain residual data, and the video data is reconstructed based on the residual data.
  • a video decoder comprising:
  • An analyzing unit configured to parse the video code stream, and obtain index information of the calculation result and the coded transform coefficient matrix
  • a determining unit configured to determine a transform coefficient matrix from the plurality of candidate transform matrices according to the index information and the intra prediction mode
  • a reconstruction unit configured to inversely transform the calculation result by using the matrix of transform coefficients to obtain residual data; and reconstruct video data according to the residual data.
  • a video data encoding method comprising:
  • the intra prediction mode selecting a set of optimal transform matrices from a plurality of candidate transform matrices according to an optimization criterion, transform-comcoding the prediction residuals to obtain a transform result;
  • the encoded code stream is generated based on the transformed result and the selected transform matrix index information.
  • a video decoding method comprising:
  • a video data encoding method comprising:
  • a video decoding method comprising:
  • a video data encoder comprising:
  • a residual generating unit configured to generate a prediction residual according to the input video data
  • a transform unit configured to perform transform coding on the prediction residual by selecting a set of optimal transform matrices from the plurality of candidate transform matrices according to an optimization criterion, to obtain a transform result
  • a code stream generating unit configured to encode the selected transform matrix index information according to the intra prediction mode according to the transform result, to generate an encoded code stream.
  • a video decoder comprising:
  • An analyzing unit configured to parse the video code stream, obtain a transform result, and obtain transform matrix index information according to an intra prediction mode
  • a determining unit configured to determine a transformation matrix from the plurality of candidate transformation matrices according to the transformation matrix index information
  • a reconstruction unit configured to inverse transform the transformation result by using the determined transformation matrix to obtain residual data; and reconstruct the video data according to the residual data.
  • the video encoding and decoding method and device may perform transform coding on a prediction residual by selecting an optimal transform matrix from a plurality of candidate transform matrices according to a rate prediction criterion according to an intra prediction mode, to obtain a transform result. .
  • the most efficient transform matrix can be selected and transformed according to the characteristics of each residual fast, thereby improving coding efficiency.
  • the transform coefficient matrix is also found from the plurality of candidate transform matrices by the transform coefficient matrix index information and the intra prediction mode, and the transform coefficient matrix is inverse-transformed to obtain residual data, thereby reconstructing the video data.
  • FIG. 1 is a flow chart of a video encoding method according to an embodiment of the present invention.
  • FIG. 2 is a flow chart of a video decoding method according to an embodiment of the present invention.
  • FIG. 3 is a schematic diagram of residual transform of a video encoding method according to an embodiment of the present invention.
  • FIG. 4 is a structural block diagram of a video encoder according to an embodiment of the present invention.
  • FIG. 5 is a structural block diagram of a video encoder according to another embodiment of the present invention.
  • FIG. 6 is a structural block diagram of a video decoder according to an embodiment of the present invention.
  • FIG. 7 is a structural block diagram of a video decoder according to another embodiment of the present invention.
  • FIG. 8 is a flow chart of still another video encoding method according to an embodiment of the present invention.
  • FIG. 9 is a flow chart of still another video decoding method according to an embodiment of the present invention.
  • FIG. 10 is a flow chart of still another video encoding method according to an embodiment of the present invention.
  • FIG. 11 is a flow chart of still another video decoding method according to an embodiment of the present invention.
  • FIG. 12 is a structural block diagram of still another video encoder according to an embodiment of the present invention.
  • FIG. 13 is a structural block diagram of still another video encoder according to an embodiment of the present invention.
  • FIG. 14 is a structural block diagram of still another video decoder according to an embodiment of the present invention.
  • FIG. 15 is a structural block diagram of still another video decoder according to an embodiment of the present invention.
  • the video data encoding method provided by the embodiment of the present invention is as shown in FIG. 1 , and the method steps include:
  • the row-column separation transform may also be adopted, that is, traversing the combination of all possible column transform matrices and row transform matrices in multiple candidate transform matrices according to the intra prediction mode, and selecting the rate-distortion cost of the matrix multiplication
  • the smallest transform combination is used as a matrix of transform coefficients, and the result of the transform is obtained.
  • the method may further include: a coefficient scanning process of scanning the transformed coefficients by selecting a set of coefficient scanning order according to the intra prediction mode and the transformation matrix index information.
  • the one with the lowest rate distortion cost after transform is selected as the optimal intra prediction mode, and the result is quantized and entropy encoded.
  • index information of the transform coefficient matrix can also be written in the encoded data.
  • the video coding method provided by the embodiment of the present invention may perform transform coding on a prediction residual by selecting an optimal transformation matrix from a plurality of candidate transformation matrices according to an intra prediction mode according to a rate distortion criterion to obtain a transformation result.
  • the most efficient transform matrix can be selected and transformed according to the characteristics of each residual fast, thereby improving the coding efficiency.
  • an intra prediction mode select a set of optimal transform matrices from a plurality of candidate transform matrices according to a rate distortion criterion, and transform and encode the prediction residuals to obtain a transform result.
  • the selected set of optimal transform matrices may be a non-separable transform matrix; or a pair of transform matrices, that is, a column transform matrix and a row transform matrix.
  • a set of optimal transform matrices is selected from a plurality of candidate transform matrices according to a rate distortion criterion to transform and encode the prediction residual, and the transform result is obtained, according to an intra prediction mode.
  • the plurality of candidate transformation matrices are used to transform and encode the prediction residuals, and a set of optimal transformation matrices are selected according to the rate distortion criterion, and the transformation results corresponding to the set of optimal transformation matrices are used for subsequent and selected transformation matrix indexes.
  • the information generates an encoded code stream.
  • the row-column separation transform may also be adopted, that is, traversing the combination of all possible column transform matrices and row transform matrices in multiple candidate transform matrices according to the intra prediction mode, and selecting the rate-distortion cost of the matrix multiplication The smallest transform combination is used as the transform matrix, and the transform result is obtained. That is According to the intra prediction mode, traversing the combination of all the column transformation matrices and the row transformation matrices in the plurality of candidate transformation matrices, selecting the transform combination with the lowest rate distortion cost after the residual transform coding as the optimal transformation matrix, and the most The transform result corresponding to the optimal transform matrix is used to generate the encoded code gram for the subsequent and selected transform matrix index information.
  • the embodiment of the present invention may further include: a coefficient scanning process of scanning the transformed coefficients by selecting a set of coefficient scanning order according to the intra prediction mode and the transformation matrix index information.
  • the one with the lowest rate distortion cost after transform is selected as the optimal intra prediction mode, and the result is quantized and entropy encoded. That is, the prediction residual is encoded in various coding manners, and the mode in which the rate loss rate is the smallest is used as the intra prediction mode, and the coding result is obtained.
  • the generating the encoded code stream according to the transformation result and the selected transformation matrix index information includes writing the transformation matrix index information into the encoded data.
  • writing the transformation matrix index information into the encoded data comprises: jointly coding the index information of a pair of transformation matrices, or index information of a pair of transformation matrices Encoding separately; writing the index information encoding result into the encoded data.
  • the joint coding indicates that the column transformation matrix and the row transformation matrix appear in pairs, and each row transformation matrix corresponds to a constraint.
  • one row transformation matrix can correspond to any one column transformation matrix, which can save the storage space of the transformation matrix.
  • the video coding method provided by the embodiment of the present invention may perform transform coding on a prediction residual by selecting an optimal transformation matrix from a plurality of candidate transformation matrices according to an intra prediction mode according to a rate distortion criterion to obtain a transformation result.
  • the most efficient transform matrix can be selected and transformed according to the characteristics of each residual block, thereby improving coding efficiency.
  • the video decoding method provided by the embodiment of the present invention is as shown in FIG. 2, and the method includes:
  • the S20K parses the video coded stream to obtain the index of the calculation result and the coded transform coefficient matrix.
  • the method further includes: an inverse coefficient scanning process, wherein the inverse coefficient sweep is performed on the transformed coefficient by selecting a set of coefficient scanning order according to the index information of the intra prediction mode and the transform coefficient matrix Description.
  • the transform coefficient matrix in the step S202 may be the row transform coefficient matrix index information and the column transform coefficient matrix index information according to the index information, and the intra prediction mode. Determined from a set of candidate row transform matrices and column transform matrices.
  • the video decoding method provided by the embodiment of the present invention can parse the video encoded code stream, obtain the calculation result and the index information of the coded transform coefficient matrix, and determine the transform coefficient matrix from the plurality of candidate transform matrices according to the index information and the intra prediction mode.
  • the transform coefficient matrix is used to inverse transform the calculation result to obtain residual data, and the video data is reconstructed according to the residual data.
  • decoding can be performed without increasing complexity. Since the coding adopts the method provided in the foregoing embodiment, the optimal transformation matrix can be selected for the characteristics of the residual, so that the entropy coding efficiency is improved, and the decoding method provided in this embodiment can effectively improve the video coding and decoding. Overall efficiency.
  • the S20K parses the video encoded code stream to obtain a calculation result and a transformation matrix index information.
  • the result of the analysis includes a transformation result, that is, the calculation result used in the embodiment of the present invention is a transformation result, and the transformation result may include a transformation coefficient matrix obtained by the transformation.
  • the embodiment of the present invention further includes: an inverse coefficient scanning process, configured to perform inverse coefficient scanning on the transformed coefficient by selecting a set of coefficient scanning order according to the intra prediction mode and the transformation matrix index information.
  • the determined transformation matrix is a set of transformation matrices
  • a set of transformation matrices may be a non-separation transformation matrix; or a pair of transformation matrices, that is, a column transformation matrix and a row transformation matrix.
  • the transform matrix in the step S202 may be based on the row transform matrix index information and the column transform matrix index information in the index information, and the intra prediction mode is from one
  • the candidate row transformation matrix and the column transformation matrix are determined.
  • a set of candidate row transform matrices and column transform matrices includes a plurality of row transform matrices and column transform matrices.
  • the video decoding method provided by the embodiment of the present invention can parse the video encoded code stream, obtain the calculation result and the transformation matrix index information, and determine the transformation matrix from the plurality of candidate transformation matrices according to the transformation matrix index information and the intra prediction mode.
  • the transformation matrix inversely transforms the calculation result to obtain residual data, and reconstructs the video data according to the residual data.
  • decoding can be performed without increasing complexity. Since the coding method uses the method provided in the foregoing embodiment, the optimal transformation matrix can be selected for the characteristics of the residual, so that the entropy coding efficiency is improved, and the decoding method provided in this embodiment can effectively improve the video coding. The overall efficiency of decoding.
  • the video data encoding method provided by the embodiment of the present invention is described by taking intraframe coding in H.264/AVC as an example.
  • a new macroblock coding mode is adopted: that is, the method provided by the embodiment of the present invention, the I4MB_RD0T, I16MB_RD0T, and I8MB_RD0T modes are used to encode the macroblock, and the corresponding rate distortion costs RDcost_I4MB_RD0T, RDcost_I16MB_RD0T, and RDcost_I8MB_RDOT 0 are calculated.
  • the specific encoding process of I4MB_RD0T, I16MB_RD0T and I8MB_RD0T is as follows. a) When encoding a macroblock using the so-called I4MB_RD0T mode, first, similar to the 14MB encoding process, a 16 x 16 size macroblock is divided into 16 4 x 4 subblocks that do not overlap each other. Then, the best prediction direction for each sub-block is selected. This step is different from the I4MB encoding process. The difference is that when transforming the residual, multiple sets of candidate transform matrices are selected according to the current intra prediction mode, and the residuals are transform-coded, and the code rate R is recorded separately.
  • the optimal prediction direction of each 16 16 size block is selected, which is different from the I16MB encoding process, and different
  • the method is: when transforming the residual, selecting a given set of candidate transform matrices according to the prediction direction, and traversing the combinations of all possible column transform matrices and row transform matrices of the set of candidate transform matrices, respectively recording
  • the code rate R and the distortion D, the rate-distortion cost, and the transformation matrix combination with the least selectivity cost penalty are selected as the optimal combination and used for the coding of the actual residual data.
  • Step 2 In the macroblock coding mode of I4MB_RD0T, I 16MB_RD0T and I8MB_RD0T, the residual after each sub-block conversion is selected according to the intra prediction mode and the coefficient scan order corresponding to the transformation matrix selection.
  • Step 3 According to the rate-distortion cost corresponding to the four intra macroblock coding modes 14MB, I16MB, I8MB, I4MB_RD0T.I 16MB-RD0T and I8MB-RD0T obtained in step 1, the mode with the lowest rate-distortion cost is selected as the best. Macroblock coding mode. If the best macroblock coding mode is I4MB, I16MB or I8MB, then when the macroblock header information is entropy encoded, the syntax element RD0T_0N is written after the syntax element CBP, and the syntax element is assigned a value of 0, indicating that no use is proposed.
  • the best macroblock mode is I4MB_RD0T, I 16MB_RD0T or I8MB_RD0T
  • the syntax element RD0T_0N is written after the syntax element CBP, and the syntax element is assigned a value of 1, indicating use Technique, and entropy coding in the sequence of the syntax element to write each block of the current macroblock The transformation matrix index number used.
  • the syntax change of the H.264 video coding standard in the embodiment of the present invention is shown in Table 1.
  • the syntax element RD0T_0N is written after the original syntax element CBP. If the macroblock mode is I4MB, I 16MB or I8MB, then RD0T_0N takes a value of 0, otherwise if the macroblock mode is I4MB_RD0T, I 16MB—RD0T or I 8MB—RD0T, then RD0T—0N takes a value of 1.
  • Transform_matr ix_ index (transform matrix index) is written after the syntax element RD0T_0N, which contains each block in the macroblock The index number information of the selected transformation matrix.
  • the brightness is analyzed below.
  • the complexity of the method provided by the embodiment of the present invention and the transform scheme of the MDDT are different in the following two points: 1)
  • the additional complexity of this part of the operation compared to the MDDT technique is the entropy decoding of the two new syntax elements: the RD0T_0N flag and the transformation matrix index number.
  • the complexity of this part is negligible for the complexity of the entire decoding process.
  • the decoder needs to select a corresponding coefficient scan order and a transform matrix according to the transformed transform matrix index number.
  • This part of the operation is the same as the MDDT technique, but requires additional storage space to store the candidate transformation matrix and the coefficient scan order. Since the I 4MB mode has 9 prediction directions, if there are 2 candidate row transformation matrices and 2 candidate column transformation matrices in each direction, the transformation matrix has each element as an integer, and the value ranges from 0 to 128.
  • the additional complexity of this part of the operation compared to the MDDT technique is the entropy encoding of the two new syntax elements: the RD0T_0N flag and the transformation matrix index number. This additional complexity is negligible for the complexity of the entire coding process.
  • the encoder needs additional storage space to be reserved Select the transformation matrix and the coefficient scan order.
  • the required storage space is the same as the decoding side, which is 18.42 KB.
  • two macroblock coding modes I4MB_RD0T, I 16MB-RD0T and I8MB_RD0T are two macroblock coding modes.
  • the encoder needs to select an optimal transform matrix for each residual block.
  • the video coding method provided by the embodiment of the present invention may perform transform coding on a prediction residual by selecting an optimal transformation matrix from a plurality of candidate transformation matrices according to an intra prediction mode according to a rate distortion criterion to obtain a transformation result.
  • the most efficient transform matrix can be selected and transformed according to the characteristics of each residual block, thereby improving coding efficiency.
  • the video data encoder provided by the embodiment of the present invention, as shown in FIG. 4, includes:
  • a residual generating unit 401 configured to generate a prediction residual according to the input video data
  • the transform unit 402 is configured to perform transform coding on the prediction residual by selecting a set of optimal transform matrices from the plurality of candidate transform matrices according to the intra prediction mode according to the intra prediction mode, to obtain a transform result;
  • the code stream generating unit 403 is configured to generate an encoded code stream according to the transform result and the selected transform matrix index information.
  • the transforming unit 402 is specifically configured to traverse a combination of all the column transformation matrices and the row transform matrices in the plurality of candidate transform matrices according to the intra prediction mode, and select a transform combination with the most rate-distortion rate distortion after the matrix multiplication. Excellent transform coefficient matrix, and get the transform result.
  • the video data encoder further includes:
  • the coefficient scanning unit 501 is specifically configured to scan the transformed coefficients by selecting a set of coefficient scanning order according to the intra prediction mode and the transformation matrix index information.
  • the determining unit 502 is configured to determine, as an intra prediction mode, a mode in which the rate residual distortion cost is minimized after encoding the prediction residual in various coding manners, and obtain an encoding result.
  • the index coding unit 503 is configured to write index information of the transform coefficient matrix into the encoded data.
  • the video encoder provided by the embodiment of the present invention may perform transform coding on a prediction residual by selecting an optimal transform matrix from a plurality of candidate transform matrices according to a rate prediction criterion according to an intra prediction mode, to obtain a transform result.
  • the video decoding method provided by the embodiment of the present invention is further described below with reference to FIG. 4 and FIG. 5.
  • the video data encoder provided by the embodiment of the present invention, as shown in FIG. 4, includes:
  • a residual generating unit 401 configured to generate a prediction residual according to the input video data
  • the transform unit 402 is configured to perform transform coding on the prediction residual by selecting a set of optimal transform matrices from the plurality of candidate transform matrices according to the intra prediction mode according to the intra prediction mode, to obtain a transform result;
  • the code stream generating unit 403 is configured to generate an encoded code stream according to the transform result and the selected transform matrix index information.
  • the selected set of optimal transform matrices may be a non-separable transform matrix; or a pair of transform matrices, that is, include a column transform matrix and a row transform matrix.
  • the transform unit 402 performs transform coding on the prediction residual by selecting an optimal transform matrix from the plurality of candidate transform matrices according to the intra prediction mode, and obtains the transform result, that is, according to the frame.
  • the intra prediction mode uses a plurality of candidate transform matrices to transform and encode the prediction residuals, selects an optimal set of transform matrices according to the rate distortion criterion, and uses the transform results corresponding to the set of optimal transform matrices for subsequent and selected purposes.
  • the transform matrix index information generates an encoded code stream.
  • the transforming unit 402 is specifically configured to: traverse a combination of all the column transformation matrices and the row transform matrices in the plurality of candidate transform matrices according to the intra prediction mode, and select a transform combination with the most rate-distortion cost after matrix multiplication as the optimal Transform the matrix and get the result of the transformation. That is, according to the intra prediction mode, traversing the combination of all the column transformation matrices and the row transformation matrices in the plurality of candidate transformation matrices, and selecting the transformation combination of the residual distortion coding and the rate distortion cost most d, as the optimal transformation matrix, and The transform result corresponding to the set of optimal transform matrices is used for subsequent and selected transform matrix index information to generate an encoded code stream.
  • the video data encoder further includes:
  • the coefficient scanning unit 501 is specifically configured to scan the transformed coefficients by selecting a set of coefficient scanning order according to the intra prediction mode and the transformation matrix index information.
  • the determining unit 502 is configured to determine, as an intra prediction mode, a mode in which the rate residual distortion cost is minimized after encoding the prediction residual in various coding manners, and obtain an encoding result.
  • the index coding unit 503 is configured to write transform matrix index information into the encoded data.
  • writing the transformation matrix index information into the encoded data comprises: jointly coding the index information of a pair of transformation matrices, or pairing a transformation moment
  • the index information of the array is separately encoded; the index information encoding result is written into the encoded data.
  • the joint coding indicates that the column transformation matrix and the row transformation matrix appear in pairs, and each row transformation matrix corresponds to a constraint.
  • one row transformation matrix can correspond to any one column transformation matrix, which can save the storage space of the transformation matrix.
  • the video encoder provided by the embodiment of the present invention may perform transform coding on the prediction residual by selecting an optimal transform matrix from a plurality of candidate transform matrices according to an intra prediction mode according to a rate distortion criterion to obtain a transform result.
  • transform coding By encoding in this way, the most efficient transform matrix can be selected and transformed according to the characteristics of each residual block, thereby improving coding efficiency.
  • the video decoder provided by the embodiment of the present invention, as shown in FIG. 6, includes:
  • the parsing unit 601 is configured to parse the video code stream to obtain the calculation result and the index information of the coded transform coefficient matrix
  • a determining unit 602 configured to determine a matrix of transform coefficients from the plurality of candidate transform matrices according to the index information and the intra prediction mode;
  • the reconstruction unit 603 is configured to inversely transform the calculation result by using a matrix of transform coefficients to obtain residual data; and reconstruct the video data according to the residual data.
  • the video decoder further includes:
  • the inverse coefficient scanning unit 701 is configured to perform inverse coefficient scanning on the transformed coefficients by selecting a set of coefficient scanning order according to the intra prediction mode and the index information of the transform coefficient matrix.
  • the video decoder provided by the embodiment of the present invention can parse the video encoded code stream to obtain the calculation result and the index information of the coded transform coefficient matrix; and determine the transform from the multiple candidate transform matrices according to the index information and the intra prediction mode.
  • the coefficient matrix is inversely transformed by the matrix of transform coefficients to obtain residual data, and the video data is reconstructed according to the residual data. In this way, decoding can be performed without increasing complexity.
  • the optimal transformation matrix can be selected for the characteristics of the residual, so that the entropy coding efficiency is improved, and the decoding method provided in this embodiment can effectively improve the video coding.
  • the video decoding method provided by the embodiment of the present invention is further described below with reference to FIG. 6 and FIG. 7 :
  • the video decoder provided by the embodiment of the present invention, as shown in FIG. 6, includes:
  • the parsing unit 601 is configured to parse the video code stream to obtain a calculation result and a transformation matrix index information
  • a determining unit 602 configured to determine a transform matrix from the plurality of candidate transform matrices according to the transform matrix index information and the intra prediction mode;
  • the reconstruction unit 603 is configured to inversely transform the calculation result by using the determined transformation matrix to obtain residual data; and reconstruct the video data according to the residual data.
  • the parsing unit 601 parses the result including the transform result, that is, the calculation result used in the embodiment of the present invention is the transform result, and the transform result may include the transformed transform coefficient matrix.
  • the determined transformation matrix is a set of transformation matrices, and a set of transformation matrices may be a non-separation transformation matrix; or a pair of transformation matrices, that is, a column transformation matrix and a row transformation matrix.
  • the determining unit 602 is configured to use the row transform matrix index information and the column transform matrix index information in the transform matrix index information, and the intra prediction mode from a set of candidate row transform matrices.
  • the transformation matrix is determined in the and column transformation matrices.
  • a set of candidate row transform matrices and column transform matrices may include a plurality of row transform matrices and column transform matrices.
  • the reconstruction unit 603 inversely transforms the calculation result by using the row transformation matrix and the column transformation matrix to obtain residual data, and reconstructs the video data based on the residual data.
  • the video decoder further includes:
  • the inverse coefficient scanning unit 701 is configured to perform inverse coefficient scanning on the transformed coefficients by selecting a set of coefficient scanning order according to the intra prediction mode and the transformation matrix index information.
  • the video decoder provided by the embodiment of the present invention is capable of parsing a video encoded code stream to obtain a calculation result and transform matrix index information; determining a transform matrix from a plurality of candidate transform matrices according to the index information and the intra prediction mode, and using the transform matrix
  • the calculation result is inversely transformed to obtain residual data, and the video data is reconstructed based on the residual data.
  • decoding can be performed without increasing complexity. Since the coding adopts the method provided by the foregoing embodiment, the optimal transformation matrix can be selected for the characteristics of the residual, so that the entropy coding efficiency is improved, and then the decoding method provided by this embodiment is used. Can effectively improve the overall efficiency of video codec.
  • An embodiment of the present invention further provides a video data encoding method. As shown in FIG. 8, the method steps include:
  • the selected set of optimal transform matrices may be a non-separable transform matrix; or a pair of transform matrices, that is, include a column transform matrix and a row transform matrix.
  • the optimization criteria can be rate distortion criteria, absolute error and (SAD of Sso of Lute Difference), coded bits or distortion.
  • the selection according to the optimization criteria may include multiple ways, such as the option rate distortion minimum, the SAD minimum, the least coded bit, or the least distortion.
  • an optimal set of transform matrices is selected from a plurality of candidate transform matrices according to an optimization criterion, and the prediction residual is transform-encoded, and the transform result is obtained, according to an intra prediction mode.
  • the plurality of candidate transformation matrices are used to transform and encode the prediction residuals, and an optimal set of transformation matrices are selected according to the optimization criterion, and the transformation results corresponding to the set of optimal transformation matrices are used for subsequent and selected transformation matrix indexes.
  • the information generates an encoded code stream.
  • a row-column separation transform may also be adopted, that is, traversing a combination of all possible column transform matrices and row transform matrices in a plurality of candidate transform matrices according to an intra prediction mode, and selecting a residual
  • the transform combination with the most cost-optimized criterion after the difference transform coding is used as the optimal transform matrix, and the transform result is obtained.
  • the intra prediction mode traversing the combination of all the column transformation matrices and the row transformation matrices in the plurality of candidate transformation matrices, selecting the transform combination with the least cost optimization criterion after the residual transform coding as the optimal transformation matrix, and The transform result corresponding to the set of optimal transform matrices is used to generate the encoded code stream for the subsequent and selected transform matrix index information.
  • the embodiment of the present invention may further include: a coefficient scanning process of scanning the transformed coefficients by selecting a set of coefficient scanning order according to the intra prediction mode and the transformation matrix index information.
  • the one with the lowest cost of the optimized optimization criterion is selected as the optimal intra prediction mode, and the result is quantized and then entropy encoded. That is, the prediction residual is encoded in various coding manners, The mode in which the optimization criterion is the least expensive is used as the intra prediction mode to obtain the coding result.
  • the transforming result and the selected transform matrix index information are used to generate an encoded code stream, and the transform matrix index information is written into the encoded data.
  • writing the transformation matrix index information into the encoded data comprises: jointly coding the index information of a pair of transformation matrices, or index information of a pair of transformation matrices Encoding separately; writing the index information encoding result into the encoded data.
  • a row transformation matrix can correspond to any column transformation matrix, which saves the storage space of the transformation matrix.
  • the video coding method provided by the embodiment of the present invention may perform transform coding on a prediction residual by selecting an optimal transformation matrix from a plurality of candidate transformation matrices according to an intra prediction mode according to an optimization criterion, to obtain a transformation result.
  • an optimal transformation matrix from a plurality of candidate transformation matrices according to an intra prediction mode according to an optimization criterion, to obtain a transformation result.
  • the embodiment of the invention further provides a video decoding method. As shown in FIG. 9, the method includes:
  • the embodiment of the present invention further includes: an inverse coefficient scanning process, configured to perform inverse coefficient scanning on the transformed coefficient by selecting a set of coefficient scanning order according to the intra prediction mode and the transformation matrix index information.
  • S902. Determine a set of transform matrices from the plurality of candidate transform matrices according to the transform matrix index information and the intra prediction mode, and inverse transform the transform result by using the set of transform matrices to obtain residual data, and reconstruct the video according to the residual data. data.
  • the determined set of transform matrices may be a non-separable transform matrix; or may be a pair of transform matrices, that is, include a column transform matrix and a row transform matrix.
  • the result of the analysis includes a transformation result, that is, the transformation result used in the embodiment of the present invention is a calculation result, and the transformation result may include a transformation coefficient matrix obtained by the transformation.
  • the set of transformation matrices may be determined from row transformation matrix index information and column transformation matrix index information in the transformation matrix index information, and intra prediction modes from a plurality of candidate row transformation matrices and column transformation matrices.
  • the video decoding method provided by the embodiment of the present invention is capable of parsing a video encoded code stream, obtaining a transform result and transform matrix index information, and determining a set of transform matrices from the plurality of candidate transform matrices according to the transform matrix index information and the intra prediction mode.
  • the transformation result is inversely transformed by the transformation matrix to obtain residual data, and the video data is reconstructed according to the residual data.
  • decoding can be performed without increasing complexity. Since the coding method uses the method provided in the foregoing embodiment, the optimal transformation matrix can be selected for the characteristics of the residual, so that the entropy coding efficiency is improved, and the decoding method provided in this embodiment can effectively improve the video coding and decoding. Overall efficiency.
  • the video data encoding method provided by the embodiment of the present invention includes: S100K generating a prediction residual according to the input video data.
  • S1002 Select a set of optimal transform matrices from the plurality of candidate transform matrices according to an optimization criterion, and transform and encode the prediction residuals to obtain a transform result.
  • the selected set of optimal transform matrices may be a non-separable transform matrix; or a pair of transform matrices, that is, include a column transform matrix and a row transform matrix.
  • the optimization criteria can be rate distortion criteria, absolute error and (SAD of Absolute Difference), coded bits or distortion.
  • the selection according to the optimization criteria may include multiple ways, such as the option rate distortion minimum, the SAD minimum, the least coded bit, or the least distortion.
  • a set of optimal transform matrices is selected from a plurality of candidate transform matrices according to an optimization criterion to transform and encode the prediction residuals, and the transform result is obtained, and the prediction residuals are performed by using multiple candidate transform matrices.
  • Transform coding selecting an optimal set of transform matrices according to an optimization criterion, and using the transform results corresponding to the set of optimal transform matrices for subsequent and selected transform matrix index information to generate an encoded code stream.
  • row-column separation transform may also be used, that is, traversing all possible combinations of column transform matrices and row transform matrices in multiple candidate transform matrices, and selecting the most residual transform coding.
  • the transformation combination with the least cost criterion is used as the optimal transformation matrix, and the transformation result is obtained. That is, traversing a combination of all column transformation matrices and row transformation matrices in a plurality of candidate transformation matrices,
  • the transform combination with the least cost optimization criterion after the residual transform coding is selected as the optimal transform matrix, and the transform result corresponding to the set of optimal transform matrices is used for the subsequent and selected transform matrix index information to generate the encoded code stream.
  • S1003 Encode the selected transform matrix index information according to the transform result according to the intra prediction mode to generate an encoded code
  • the embodiment of the present invention may further include: a coefficient scanning process of scanning the transformed coefficients by selecting a set of coefficient scanning order according to the transformation matrix index information.
  • the one with the least cost of the optimized optimization criterion is selected as the optimal intra prediction mode, and the result is quantized and then entropy encoded. That is, the prediction residual is encoded in various coding manners, and the mode in which the optimization criterion cost is the smallest is used as the intra prediction mode, and the coding result is obtained.
  • the selected transformation matrix index information is encoded according to the intra prediction mode, and the generated coded stream includes selecting a transformation matrix index according to the selected intra prediction mode.
  • the encoding method of the information the conversion matrix index information is written into the encoded data.
  • different transform matrix index information encoding methods may be used to write transform matrix index information into the encoded data.
  • writing the transformation matrix index information into the encoded data comprises: jointly coding the index information of a pair of transformation matrices, or index information of a pair of transformation matrices Encoding is performed separately; the index information encoding result is written into the encoded data according to the intra prediction mode.
  • a row transformation matrix can correspond to any column transformation matrix, which saves the storage space of the transformation matrix.
  • an optimal set of transform matrices may be selected from a plurality of candidate transform matrices according to an optimization criterion to transform and encode the prediction residuals to obtain a transform result.
  • the video decoding method provided by the embodiment of the present invention, as shown in FIG. 11, the method includes:
  • S110 parses the video encoded code stream, obtains a transform result, and obtains according to an intra prediction mode. To transform matrix index information.
  • the result of the analysis includes a transformation result, that is, the transformation result used in the embodiment of the present invention is a calculation result, and the transformation result may include a transformation coefficient matrix obtained by the transformation.
  • Obtaining the transformation matrix index information according to the intra prediction mode includes: selecting a decoding method of the transformation matrix index information according to the intra prediction mode, and decoding the transformation matrix index information. For different intra prediction modes, different parsing methods can be used to parse the video bitstream to obtain transform matrix index information.
  • the embodiment of the present invention further includes: an inverse coefficient scanning process, that is, performing inverse coefficient scanning on the transformed coefficient by selecting a set of coefficient scanning order according to the transformation matrix index information.
  • S1102 Determine a transformation matrix from the plurality of candidate transformation matrices according to the transformation matrix index information, and inversely transform the transformation result by using the determined transformation matrix to obtain residual data, and reconstruct the video data according to the residual data.
  • the determined transformation matrix may be a set of transformation matrices, and a set of transformation matrices may be a non-separation transformation matrix; or a pair of transformation matrices, that is, a column transformation matrix and a row transformation matrix.
  • the transform matrix in step S1102 may be a row transformation matrix from a set of candidate rows according to row transform matrix index information and column transform matrix index information in the index information. Determined in the column and column transformation matrix.
  • a set of candidate row transform matrices and column transform matrices includes a plurality of row transform matrices and column transform matrices.
  • the video decoding method provided by the embodiment of the present invention can parse the video encoded code stream, obtain a transform result, and parse the transform matrix index information according to the intra prediction mode, and determine the transform from the multiple candidate transform matrices according to the transform matrix index information.
  • the matrix transforms the transform result by the transform matrix to obtain residual data, and reconstructs the video data according to the residual data.
  • decoding can be performed without increasing complexity. Since the coding adopts the method provided in the foregoing embodiment, the optimal transformation matrix can be selected for the characteristics of the residual, so that the entropy coding efficiency is improved, and the decoding method provided in this embodiment can effectively improve the video coding and decoding. Overall efficiency.
  • the video data encoder provided by the embodiment of the present invention, as shown in FIG. 12, includes:
  • a residual generating unit 1201 configured to generate a prediction residual according to the input video data;
  • the transform unit 1202 transforming and encoding the prediction residuals by selecting a set of optimal transform matrices from the plurality of candidate transform matrices according to an optimization criterion, to obtain a transform result;
  • the code stream generating unit 1203 is configured to encode the selected transform matrix index information according to the transform result according to the intra prediction mode to generate an encoded code stream.
  • the selected set of optimal transform matrices may be a non-separable transform matrix; or a pair of transform matrices, that is, include a column transform matrix and a row transform matrix.
  • the optimization criteria described include: rate distortion criteria, absolute error and SAD, coded bits or distortion
  • the transforming unit 1202 is specifically configured to traverse a combination of all the column transform matrices and the row transform matrices in the plurality of candidate transform matrices, and select a transform combination with the least cost optimization criterion after the residual transform encoding as the optimal transform. Matrix, and get the result of the transformation.
  • the code stream generating unit 1203 encodes the selected transform matrix index information according to the intra-prediction mode according to the transform result, and generates the encoded code stream according to the selected intra-frame.
  • the prediction mode an encoding method of transforming matrix index information is selected, and the transform matrix index information is written into the encoded data.
  • the video data encoder further includes:
  • the coefficient scanning unit 1301 is configured to select a set of coefficient scanning order to scan the transformed coefficients according to the transformation matrix index information.
  • the determining unit 1302 is configured to determine, as an intra prediction mode, a mode in which the cost of the optimization criterion is minimized after encoding the prediction residual in various coding manners, and obtain an encoding result.
  • the index coding unit 1303 is configured to select an encoding method of transforming matrix index information according to the selected intra prediction mode, and write the transform matrix index information into the encoded data.
  • the set of optimal transformation matrices is a pair of transformation matrices, selecting an encoding method of the transformation matrix index information according to the selected intra prediction mode, and writing the transformation matrix index information into the encoded data.
  • the method comprises: jointly coding the index information of a pair of transformation matrices, or respectively encoding the index information of the pair of transformation matrices; and selecting an encoding method of the transformation matrix index information according to the selected intra prediction mode, The transformation matrix index information is written in the encoded data.
  • a corresponding column transformation matrix; respectively coding indicates that the column transformation matrix and the row transformation matrix have no correspondence Restrictions, for example, a row transformation matrix can correspond to any one column transformation matrix, which can save the storage space of the transformation matrix.
  • the video encoder provided by the embodiment of the present invention may perform transform coding on the prediction residual by selecting an optimal transform matrix from a plurality of candidate transform matrices according to an optimization criterion to obtain a transform result.
  • an optimal transform matrix from a plurality of candidate transform matrices according to an optimization criterion to obtain a transform result.
  • the video encoder provided by the embodiment of the present invention may perform transform coding on the prediction residual by selecting an optimal transform matrix from a plurality of candidate transform matrices according to an intra prediction mode according to a rate distortion criterion to obtain a transform result.
  • transform coding By encoding in this way, the most efficient transform matrix can be selected and transformed according to the characteristics of each residual block, thereby improving coding efficiency.
  • the video decoder provided by the embodiment of the present invention, as shown in FIG. 14, includes:
  • the parsing unit 1401 is configured to parse the video code stream, obtain a transform result, and obtain transform matrix index information according to the intra prediction mode;
  • a determining unit 1402 configured to determine a transform matrix from the plurality of candidate transform matrices according to the transform matrix index information
  • the reconstruction unit 1403 is configured to inverse transform the transformation result by using the determined transformation matrix to obtain residual data; and reconstruct the video data according to the residual data.
  • the determined transformation matrix is a set of transformation matrices, and a set of transformation matrices may be a non-separation transformation matrix; or a pair of transformation matrices, that is, a column transformation matrix and a row transformation matrix.
  • the result of the analysis includes a transformation result, that is, the transformation result used in the embodiment of the present invention is a calculation result, and the transformation result may include a transformation coefficient matrix obtained by the transformation.
  • the parsing unit 1401 obtains the transform matrix index information according to the intra prediction mode, and: according to the intra prediction mode, selecting a decoding method of the transform matrix index information, and decoding the transform matrix index information.
  • the video decoder further includes:
  • An inverse coefficient scanning unit 1501 configured to select a group of coefficient scans according to the transformation matrix index information The inverse coefficients are scanned sequentially for the transformed coefficients.
  • the video decoder provided by the embodiment of the present invention can parse the video encoded code stream to obtain a transform result, and parse the transform matrix index information according to the intra prediction mode, and determine the transform from the multiple candidate transform matrices according to the transform matrix index information.
  • the matrix transforms the transform result by the transform matrix to obtain residual data, and reconstructs the video data according to the residual data.
  • decoding can be performed without increasing complexity. Since the coding adopts the method provided in the foregoing embodiment, the optimal transformation matrix can be selected for the characteristics of the residual, so that the entropy coding efficiency is improved, and the decoding method provided in this embodiment can effectively improve the video coding and decoding. Overall efficiency.
  • the foregoing storage medium includes: a medium that can store program codes, such as a ROM, a RAM, a magnetic disk, or an optical disk.

Landscapes

  • Engineering & Computer Science (AREA)
  • Multimedia (AREA)
  • Signal Processing (AREA)
  • Physics & Mathematics (AREA)
  • Discrete Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Compression Or Coding Systems Of Tv Signals (AREA)

Description

一种视频编解码方法及设备 本申请要求了 2009年 10月 23日提交的, 申请号为 200910209013. 9, 发明 名称为 "一种视频编解码方法及设备" , 2010 年 4 月 9 日提交的, 申请号为 201010147581. 3, 发明名称为 "一种视频编解码方法及设备" 和 2010年 6月 17 日提交的, 申请号为 201010213791. 8 , 发明名称为 "一种视频编解码方法及设 备" 的中国专利申请的优先权, 其全部内容通过引用结合在本申请中。
技术领域
本发明涉及通信领域, 尤其涉及一种视频编解码方法及设备。 背景技术
一个完整的视频编解码系统包括编码器与解码器两部分。 大致而言, 在混 合编码框架下的编码端, 视频信号首先会经过预测模块, 编码器依照一定的最 优化准则从若干种预测模式中选择最佳的一种, 然后生成残差信号; 残差信号 经过变换、 量化后进入熵编码模块, 并最终形成输出码流。 在解码端, 首先从 码流中解析出预测模式信息, 生成与编码端完全一致的预测信号; 接着解析出 码流中已经量化过的变换系数值, 进行反量化与反变换, 生成重构残差信号; 最后用预测信号与重建残差信号合成出重构视频信号。
在混合编码框架下, 编码的流程当中包含一项关键的技术: 变换。 变换的 作用是通过对残差块进行某种线性运算, 将残差变换为另外一种表达形式, 并 且在这种表达形式下, 数据的能量集中在少数的几个变换系数上, 其余大部分 的系数的能量很低或者为零, 通过这样的变换, 能够使后续的熵编码高效地进 行。 在视频编码中, 对于某个残差块 X而言, 如果将 X看作为矩阵, 那么变换 实际上就是进行矩阵相乘, 相乘的一种形式为 F=C · X · R , 其中, C和 R是和 X 尺寸相同的变换矩阵, F 是变换得到的变换系数矩阵。 由于离散余弦变换 ( Di s crete Cos ine Transform, DCT )在复杂度和性能这两方面, 相对于其他 现有的变换而言有更好的折中, 因此, 被广泛釆纳。
在视频编码技术中, 一项被称为依赖方向变换模式 (Mode dependent Di rect iona l Transform, MDDT ) 的技术被釆纳。 其核心思想是: ①由于不同帧 内预测模式得到的残差体现着不同的统计特性 , 所以变换应该根据预测方向的 不同, 采用不同的变换矩阵来提高压缩编码效率, ②为了降低变换的复杂度, MDDT采用行列分离的变换形式, 得到一对变换矩阵, 即一个行列变换矩阵 Ci和 —个行变换矩阵 Ri , 那么变换的过程即为 '=6 . 其中, i为对应的帧 内预测模式, J为预测残差, W为转换后的预测残差, 67和^ 7可以看到, 水平 和垂直变换由 和 两个矩阵分离开来, 这也就是所谓的行列分离的变换。
在实现上述变换的过程中, 发明人发现现有技术中至少存在如下问题: 虽然 MDDT技术能够针对帧内编码, 对不同的预测方向, 釆用不同组的变换 矩阵, 但在实际的编码过程中, 即使在一个相同的帧内预测模式下, 残差数据 的统计特性仍然会存在明显的差异, 所以上述一种帧内预测模式对应一组变换 矩阵的方法仍然不够准确 , 使得后续编码效率较低。 发明内容
本发明的实施例提供一种视频编解码方法及设备, 能够根据每个残差块的 特性, 针对性地选择有效的变换矩阵进行变换, 从而提高编码效率。
为达到上述目的, 本发明的实施例采用如下技术方案:
一种视频数据编码方法, 包括:
根据输入的视频数据生成预测残差;
根据帧内预测模式, 根据率失真准则从多个候选变换矩阵中选择一组最优 的变换矩阵对预测残差进行变换编码, 得到变换结果;
才艮据所述变换结果和所选用的变换矩阵索引信息, 生成编码码流。
一种视频数据编码器, 包括:
残差生成单元, 用于根据输入的视频数据生成预测残差;
变换单元, 用于根据帧内预测模式, 根据率失真准则从多个候选变换矩阵 中选择一组最优的变换矩阵对预测残差进行变换编码, 得到变换结果;
码流生成单元, 用于才艮据所述变换结果和所选用的变换矩阵索引信息, 生 成编码码流。 一种视频数据解码方法, 包括:
对视频编码码流进行解析, 得到计算结果和编码变换系数矩阵的索引信息; 根据所述索引信息和帧内预测模式从多个候选变换矩阵中确定变换系数矩 阵, 利用所述变换系数矩阵对所述计算结果进行反变换, 得到残差数据, 根据 所述残差数据重建视频数据。
一种视频解码器, 包括:
解析单元, 用于对视频码流进行解析, 得到计算结果和编码变换系数矩阵 的索引信息;
确定单元, 用于根据所述索引信息和所述帧内预测模式从多个候选变换矩 阵中确定变换系数矩阵;
重建单元, 用于利用所述变换系数矩阵对所述计算结果进行反变换, 得到 残差数据; 根据所述残差数据重建视频数据。
一种视频数据编码方法, 其特征在于, 包括:
根据输入的视频数据生成预测残差;
根据帧内预测模式, 根据最优化准则从多个候选变换矩阵中选择一組最优 的变换矩阵对预测残差进行变换编码, 得到变换结果;
才艮据所述变换结果和所选用的变换矩阵索引信息, 生成编码码流。
一种视频解码方法, 其特征在于, 包括:
对视频编码码流进行解析, 得到变换结果和变换矩阵索引信息;
根据所述变换矩阵索引信息和帧内预测模式从多个候选变换矩阵中确定一 组变换矩阵, 利用所述一组变换矩阵对所述变换结果进行反变换, 得到残差数 据, 根据所述残差数据重建视频数据。
一种视频数据编码方法, 其特征在于, 包括:
根据输入的视频数据生成预测残差;
根据最优化准则从多个候选变换矩阵中选择一组最优的变换矩阵对预测残 差进行变换编码, 得到变换结果;
根据所述变换结果,并根据帧内预测模式对所选用的变换矩阵索引信息进 行编码, 生成编码码流。 一种视频解码方法, 其特征在于, 包括:
对视频编码码流进行解析, 得到变换结果, 并根据帧内预测模式得到变换 矩阵索引信息;
根据所述变换矩阵索引信息从多个候选变换矩阵中确定变换矩阵 , 利用确 定的变换矩阵对所述变换结果进行反变换, 得到残差数据, 根据所述残差数据 重建视频数据。
一种视频数据编码器, 其特征在于, 包括:
残差生成单元, 用于根据输入的视频数据生成预测残差;
变换单元, 用于根据最优化准则从多个候选变换矩阵中选择一组最优的变 换矩阵对预测残差进行变换编码, 得到变换结果;
码流生成单元, 用于根据所述变换结果, 并根据帧内预测模式对所选用的 变换矩阵索引信息进行编码, 生成编码码流。
一种视频解码器, 其特征在于, 包括:
解析单元, 用于对视频码流进行解析, 得到变换结果, 并根据帧内预测模 式得到变换矩阵索引信息;
确定单元 , 用于根据所述变换矩阵索引信息从多个候选变换矩阵中确定变 换矩阵;
重建单元, 用于利用确定的变换矩阵对所述变换结果进行反变换, 得到残 差数据; 根据所述残差数据重建视频数据。
本发明实施例提供的视频编解码方法及设备, 可以根据帧内预测模式, 根 据率失真准则从多个候选变换矩阵中选择一组最优的变换矩阵对预测残差进行 变换编码, 得到变换结果。 通过这种方式进行编码, 可以根据每个残差快的特 性, 针对性地选择最有效的变换矩阵进行变换, 从而提高编码效率。 同样, 通 过变换系数矩阵索引信息和帧内预测模式也从多个候选变换矩阵中找到所述变 换系数矩阵, 利用该变换系数矩阵进行反变换, 得到残差数据, 从而重建视频 数据。 附图说明 为了更清楚地说明本发明实施例中的技术方案, 下面将对实施例或现有技 术描述中所需要使用的附图作简单地介绍, 显而易见地, 下面描述中的附图仅 仅是本发明的一些实施例, 对于本领域普通技术人员来讲, 在不付出创造性劳 动的前提下, 还可以才艮据这些附图获得其他的附图。
图 1为本发明实施例提供的视频编码方法的流程框图;
图 2为本发明实施例提供的视频解码方法的流程框图;
图 3为本发明实施例提供的视频编码方法的残差变换示意图;
图 4为本发明实施例提供的视频编码器的结构框图;
图 5为本发明另一实施例提供的视频编码器的结构框图;
图 6为本发明实施例提供的视频解码器的结构框图;
图 7为本发明另一实施例提供的视频解码器的结构框图;
图 8为本发明实施例提供的又一个视频编码方法的流程框图;
图 9为本发明实施例提供的又一个视频解码方法的流程框图;
图 10为本发明实施例提供的再一个视频编码方法的流程框图;
图 11为本发明实施例提供的再一个视频解码方法的流程框图;
图 12为本发明实施例提供的又一个视频编码器的结构框图;
图 13为本发明实施例提供的再一个视频编码器的结构框图;
图 14为本发明实施例提供的又一个视频解码器的结构框图;
图 15为本发明实施例提供的再一个视频解码器的结构框图。
具体实施方式
下面将结合本发明实施例中的附图, 对本发明实施例中的技术方案进行清 楚、 完整地描述, 显然, 所描述的实施例是本发明一部分实施例, 而不是全部 的实施例。 基于本发明中的实施例, 本领域普通技术人员在没有作出创造性劳 动前提下所获得的所有其他实施例, 都属于本发明保护的范围。
本发明实施例提供的视频数据编码方法, 如图 1所示, 该方法步驟包括:
5101、 根据输入的视频数据生成预测残差。
5102、 根据帧内预测模式, 根据率失真准则从多个候选变换矩阵中选择一 组最优的变换矩阵对预测残差进行变换编码, 得到变换结果。
在进行变换的过程中, 还可以采用行列分离变换, 即根据帧内预测模式, 遍历多个候选变换矩阵中的所有可能的列变换矩阵和行变换矩阵的组合, 选择 矩阵相乘后率失真代价最小的变换组合作为变换系数矩阵, 并得到变换结果。
S103、 才艮据变换结果和所选用的变换矩阵索引信息, 生成编码码流。
进一步地, 该方法还可以包括: 系数扫描过程, 为根据帧内预测模式和变 换矩阵索引信息选择一组系数扫描顺序对变换后的系数进行扫描。
然后选择变换后率失真代价最小的一种作为最佳帧内预测模式, 对其结果 量化后进行熵编码。
此外, 还可以将变换系数矩阵的索引信息写入编码数据中。
本发明实施例提供的视频编码方法, 可以根据帧内预测模式, 根据率失真 准则从多个候选变换矩阵中选择一组最优的变换矩阵对预测残差进行变换编 码, 得到变换结果。 通过这种方式进行编码, 可以根据每个残差快的特性, 针 对性地选择最有效的变换矩阵进行变换, 从而提高编码效率。
下面结合图 1 , 对本发明实施例提供的视频数据编码方法进行进一步说明:
5101、 根据输入的视频数据生成预测残差。
5102、 根据帧内预测模式, 根据率失真准则从多个候选变换矩阵中选择一 组最优的变换矩阵对预测残差进行变换编码, 得到变换结果。
本发明实施例中 , 选择的一组最优的变换矩阵可以为一个非分离变换矩阵; 也可以为一对变换矩阵, 即包括一个列变换矩阵和一个行变换矩阵。
本发明实施例中, 根据帧内预测模式, 根据率失真准则从多个候选变换矩 阵中选择一组最优的变换矩阵对预测残差进行变换编码 , 得到变换结果即为 , 根据帧内预测模式, 利用多个候选变换矩阵对预测残差进行变换编码, 根据率 失真准则选择一组最优的变换矩阵, 并将这组最优变换矩阵对应的变换结果用 于后续和所选用的变换矩阵索引信息生成编码码流。
在进行变换的过程中, 还可以采用行列分离变换, 即根据帧内预测模式, 遍历多个候选变换矩阵中的所有可能的列变换矩阵和行变换矩阵的组合, 选择 矩阵相乘后率失真代价最小的变换组合作为变换矩阵, 并得到变换结果。 也即 根据帧内预测模式, 遍历多个候选变换矩阵中的所有列变换矩阵和行变换矩阵 的组合, 选择残差变换编码后率失真代价最小的变换组合作为最优的变换矩阵, 并将这组最优变换矩阵对应的变换结果用于后续和所选用的变换矩阵索引信息 生成编码码克。
S103、 才艮据变换结果和所选用的变换矩阵索引信息, 生成编码码流。
进一步地, 本发明实施例还可以包括: 系数扫描过程, 为根据帧内预测模 式和变换矩阵索引信息选择一组系数扫描顺序对变换后的系数进行扫描。
然后选择变换后率失真代价最小的一种作为最佳帧内预测模式, 对其结果 量化后进行熵编码。 即以各种编码方式对所述预测残差进行编码, 以其中率失 真代价最小的模式作为帧内预测模式, 得到编码结果。
本发明实施例中 , 所述才艮据所述变换结果和所选用的变换矩阵索引信息, 生成编码码流, 包括将变换矩阵索引信息写入编码数据中。
若所述一组最优的变换矩阵为一对变换矩阵, 则将变换矩阵索引信息写入 编码数据中包括: 对一对变换矩阵的索引信息进行联合编码, 或对一对变换矩 阵的索引信息分别进行编码; 将索引信息编码结果写入编码数据中。
联合编码表明列变换矩阵和行变换矩阵成对出现 , 每一个行变换矩阵对应 限制, 例如一个行变换矩阵可以对应任意个一个列变换矩阵, 这样可以节省变 换矩阵的存储空间。
本发明实施例提供的视频编码方法, 可以根据帧内预测模式, 根据率失真 准则从多个候选变换矩阵中选择一组最优的变换矩阵对预测残差进行变换编 码, 得到变换结果。 通过这种方式进行编码, 可以根据每个残差块的特性, 针 对性地选择最有效的变换矩阵进行变换, 从而提高编码效率。
本发明实施例提供的视频解码方法, 如图 2所示, 该方法包括:
S20K 对视频编码码流进行解析, 得到计算结果和编码变换系数矩阵的索 引信息。
进一步地, 该方法还包括: 反系数扫描过程, 为根据帧内预测模式和变换 系数矩阵的索引信息, 选择一组系数扫描顺序对该变换后的系数进行反系数扫 描。
S202、 根据索引信息和帧内预测模式从多个候选变换矩阵中确定变换系数 矩阵, 利用变换系数矩阵对计算结果进行反变换, 得到残差数据, 根据残差数 据重建视频数据。
具体的, 当编码变换过程中, 采用的是分离变换时, 本步驟 S202中的变换 系数矩阵可以为根据索引信息中的行变换系数矩阵索引信息和列变换系数矩阵 索引信息, 以及帧内预测模式从一组候选的行变换矩阵和列变换矩阵中确定的。
本发明实施例提供的视频解码方法, 能够对视频编码码流进行解析, 得到 计算结果和编码变换系数矩阵的索引信息, 根据索引信息和帧内预测模式从多 个候选变换矩阵中确定变换系数矩阵, 利用变换系数矩阵对计算结果进行反变 换, 得到残差数据, 根据残差数据重建视频数据。 这样, 可以再不增加复杂度 的情况下, 进行解码。 由于该编码采用了上述实施例提供的方法, 能够针对残 差的特性选择最优的变换矩阵, 从而使熵编码效率有所提高, 再通过本实施例 提供的解码方法 , 能够有效提高视频编解码的整体效率。
下面结合图 2 , 对本发明实施例提供的视频解码方法进行进一步说明:
S20K 对视频编码码流进行解析, 得到计算结果和变换矩阵索引信息。 本发明实施例中, 解析得到的结果包括变换结果, 也即本发明实施例中用 到的计算结果即为变换结果, 变换结果可以包括经过变换后得到的变换系数矩 阵。
进一步地, 本发明实施例还包括: 反系数扫描过程, 为根据帧内预测模式 和变换矩阵索引信息, 选择一组系数扫描顺序对该变换后的系数进行反系数扫 描。
S202、 根据变换矩阵索引信息和帧内预测模式从多个候选变换矩阵中确定 变换矩阵, 利用确定的变换矩阵对计算结果进行反变换, 得到残差数据, 根据 残差数据重建视频数据。
本发明实施例中, 确定的变换矩阵是一组变换矩阵, 一组变换矩阵可以为 一个非分离变换矩阵; 也可以为一对变换矩阵, 即包括一个列变换矩阵和一个 行变换矩阵。 具体的, 当编码变换过程中, 釆用的是分离变换时, 本步驟 S202中的变换 矩阵可以为根据索引信息中的行变换矩阵索引信息和列变换矩阵索引信息, 以 及帧内预测模式从一组候选的行变换矩阵和列变换矩阵中确定的。 此处一组候 选的行变换矩阵和列变换矩阵包括多个行变换矩阵和列变换矩阵。
本发明实施例提供的视频解码方法, 能够对视频编码码流进行解析, 得到 计算结果和变换矩阵索引信息 , 根据变换矩阵索引信息和帧内预测模式从多个 候选变换矩阵中确定变换矩阵, 利用变换矩阵对计算结果进行反变换, 得到残 差数据, 根据残差数据重建视频数据。 这样, 可以在不增加复杂度的情况下, 进行解码。 由于该编码釆用了上述实施例提供的方法, 能够针对残差的特性选 择最优的变换矩阵, 从而使熵编码效率有所提高, 再通过本实施例提供的解码 方法, 能够有效提高视频编解码的整体效率。
本发明实施例提供的视频数据编码方法, 以 H.264/AVC 中的帧内编码为例 进行说明。
步骤 1、 在 H.264/AVC 中的帧内编码过程中, 对于每一个宏块, 首先采用 原有的 I4MB模式、 116MB模式和 18MB对宏块进行编码, 并记录其码率分别为 R_I4MB 、 R—I16MB和 R—I8MB, 失真分别为 D_I4MB、 D—I16MB和 D—I8MB; 然后分 别 计 算 率 失 真 代 价 RDcost_I4MB=D_I4MB+A · R.I4MB 、 RDcost_I16MB=D_I16MB+A · R—I16MB 和 RDcost_I8MB=D_I16MB+ λ · R.I8MB, 其中 λ 为编码过程中指定的一个常数。 之后再采用新的宏块编码模式: 即本发 明实施例提供的方法, 支设为 I4MB_RD0T、 I16MB_RD0T和 I8MB_RD0T模式对宏 块进行编码,并计算对应的率失真代价 RDcost_I4MB_RD0T、 RDcost_I16MB_RD0T 和 RDcost_I8MB_RDOT0
其中 I4MB_RD0T、 I16MB_RD0T和 I8MB_RD0T的具体编码过程如下描述。 a)、 在对宏块采用所谓的 I4MB_RD0T模式进行编码时, 首先, 类似 14MB的 编码过程, 16 x 16大小的宏块被划分为互不交叠的 16个 4 x 4的子块。 然后, 对每个子块的最佳预测方向进行选择。 这一步与 I4MB编码过程的不同, 不同之 处在于, 对残差进行变换时, 根据当前的帧内预测模式选定多组待选变换矩阵, 对残差进行变换编码, 分别记录其码率 R和失真 D, 计算率失真代价, 选择率失 真代价最小的变换矩阵组合作为最佳组合, 并用于实际的残差数据的编码。 其 残差变换过程可参见图 3。 其中, 为预测残差, 为变换后的预测残差, Cf-"和 ϋΓ."为预测方向所对应的候选变换矩阵。
b)、 在对宏块采用所谓的 I8MB_RD0T模式进行编码时, 首先, 类似 18MB的 编码过程, 16 x 16大小的宏块被划分为互不交叠的 4个 8 x 8的子块。 然后, 对 每个子块的最佳预测方向进行选择。 这一步与 18MB编码过程的不同, 不同之处 在于, 对残差进行变换时, 根据当前的帧内预测模式选定多组待选变换矩阵, 对残差进行变换编码, 分别记录其码率 R和失真 D, 计算率失真代价, 选择率失 真代价最小的变换矩阵组合作为最佳组合, 并用于实际的残差数据的编码。 其 残差变换过程可参见图 6。 其中, 为预测残差, 为变换后的预测残差, crK"和 w1为预测方向所对应的候选变换矩阵。
c)、 在对宏块采用所谓的 I 16MB_RD0T模式进行编码时, 类似 I16MB的编码 过程, 对每个 16 16大小的块的最佳预测方向进行选择, 这一步与 I16MB编码 过程的不同, 不同之处在于, 对残差进行变换时, 根据预测方向选定一组给定 的待选变换矩阵, 并遍历这組待选变换矩阵的所有可能的列变换矩阵和行变换 矩阵的组合, 分别记录其码率 R和失真 D, 计算率失真代价, 选择率失真代价最 小的变换矩阵组合作为最佳组合, 并用于实际的残差数据的编码。
步驟 2、 在 I4MB_RD0T、 I 16MB_RD0T和 I8MB_RD0T的宏块编码模式时, 对 每个子块转换后的残差根据帧内预测模式以及和变换矩阵选择所对应的系数扫 描顺序。
步骤 3、 根据步骤 1中得到的四个帧内宏块编码模式 14MB, I16MB, I8MB, I4MB_RD0T. I 16MB-RD0T 和 I8MB-RD0T所对应的率失真代价, 选择率失真代价 最小的模式作为最佳的宏块编码模式。如果最佳宏块编码模式为 I4MB、 I16MB或 者是 I8MB, 那么在对宏块头信息进行熵编码时, 在语法元素 CBP之后写入语法 元素 RD0T_0N, 并且将该语法元素赋值为 0, 表示不使用提出的技术; 如果最佳 宏块模式为 I4MB_RD0T、 I 16MB_RD0T或者是 I8MB_RD0T, 那么在宏块头信息进行 熵编码时, 在语法元素 CBP之后写入语法元素 RD0T_0N, 并且将语法元素赋值为 1 , 表示使用提出的技术, 并在该语法元素之后依次熵编码写入当前宏块各个块 所使用的变换矩阵索引号。
具体的, 本发明实施例对 H. 264视频编码标准的语法变更如表 1所示。 在 每个宏块头中, 在原有的语法元素 CBP之后写入语法元素 RD0T_0N,如果该宏块 模式为 I4MB、 I 16MB或者是 I8MB , 那么 RD0T_0N取值为 0, 否则如果该宏块 模式为 I4MB_RD0T、 I 16MB—RD0T或者是 I 8MB—RD0T, 那么 RD0T—0N取值为 1。 如 果 RD0T_0N 取值为 1 , 也就是宏块模式为 I4MB_RD0T、 I 16MB_RD0T 或者是 I8MB_RD0T , 那 么 在 语 法 元 素 RD0T_0N 之 后 写 入语 法 元 素 Transform_matr ix_ index (变换矩阵索引 ), 该语法元素包含宏块内每个块所选 中的变换矩阵的索引号信息。
Figure imgf000013_0001
Figure imgf000014_0001
表 1、 I4MB、 I8MB和 16MB模式的语法元素。
最后, 使用 KTA2.4为平台, 并采用以下设置: 全 I帧编码, CABAC, 每个 序列都测试 4个 QP点, 分别为 22, 27, 32, 37。 对比采用本发明实施例提供的 方法的变换与釆用现有 MDDT的变换方案的编码性能, 计算平均 APSNR。
QCIF序列测得的结果如表 2所示。
Figure imgf000014_0002
Foreman QCIF 0. 083
Hal l QCIF 0. 2408
Mother QCIF 0. 0519
Si lent QCIF 0. 1113
Par is QCIF 0. 2400
表 2、 QCIF序列测得的结果
CIF序列测得的结果如表 3所示。
序列 格式 PSNR (dB)
Flower CIF 0. 2596
Mobi le CIF 0. 3146
Pari s CIF 0. 1717
Stef an SIF 0. 2767
Bus CIF 0. 2398
Coas tguard CIF 0. 1469
Container CIF 0. 1911
Footbal 1 CIF 0. 1017
Foreman CIF 0. 0740
Hal l CIF 0. 2123
Si lent CIF 0. 0900
Tempete CIF 0. 1070
表 3、 CIF序列测得的结果
通过上表可以看出 ,本发明实.
有明显的性能提高。
下面, 针对本发明实施例提供的方法的复杂度进行分析说明。
下面对亮度进行分析。
在解码端, 本发明实施例提供的方法和 MDDT的变换方案的复杂度不同之处 在于以下两点: 1 )对于本发明实施例提供的方法而言, 解码器需要对每个宏块头中新增的 语法元素 RD0T_0N进行熵解码, 如果 RD0T_0N=1 , 那么解码器还需要从宏块头信 息中解码得到宏块内每个块所使用的变换矩阵的索引号。
这部分运算和 MDDT技术相比, 附加的复杂度就是对新增的两个语法元素: RD0T_0N标志和变换矩阵索引号的熵解码。 然而, 这部分的复杂度对于整个解码 过程的复杂度而言, 是可以忽略的。
2 )对于釆用了本发明实施例提供的方法的宏块( RD0T_0N=1 ), 解码器需要 根据解码得到的变换矩阵索引号选择对应的系数扫描顺序和变换矩阵。
这部分运算和 MDDT技术相比, 计算复杂度相同, 但是需要额外的存储空间 保存待选变换矩阵和系数扫描顺序。 由于 I 4MB模式有 9个预测方向, 每个方向 如果有 2个待选的行变换矩阵和 2个待选的列变换矩阵, 变换矩阵得每个元素 为整数, 取值在 0 ~ 128之间, 那么总共需要 9 X (2+2) X 16 x 7=4032比特的存 储空间; 由于 I 8MB模式有 9个预测方向, 每个方向如果有 4个待选的行变换矩 阵和 4个待选的列变换矩阵,变换矩阵得每个元素为整数,取值在 0 ~ 128之间 , 那么总共需要 9 (4+4) 64 7= 32256比特的存储空间; 同样地, 对于 I 16MB 模式, 总共有 4个预测方向, 每个方向如果有 8个待选的行变换矩阵和 8个待 选的列变换矩阵, 那么总共需要 4 X (8+8) 256 7=114688比特的存储空间; 因此, I 4MB、 I 16MB和 Ι 8ΜΒ合起来总共需要 150976比特, 也就是 18. 42KB的存 储空间。 另外, 由于记录系数扫描顺序的数组所占用的空间远远小于变换矩阵 所占用的空间, 这里不再赘述讨论。
在编码端, 本方法和 MDDT的变换方案的复杂度不同之处在于以下三点:
1 )对于本发明实施例提供的方法而言, 编码器需要在每个宏块的宏块头 信息中熵编码写入新增的语法元素 RD0T_0N, 如果 RD0T_0N=1 , 那么编码器还需 信息中。 这部分运算和 MDDT技术相比, 附加的复杂度就是对新增的两个语法元 素: RD0T_0N标志和变换矩阵索引号的熵编码。这部分附加的复杂度对于整个编 码过程的复杂度而言, 是可以忽略的。
2 )对于本发明实施例提供的方法而言, 编器需要额外的存储空间保存待 选变换矩阵和系数扫描顺序。 所需的存储空间大小和解码端相同, 为 18. 42KB。 本发明实施例的方法对于帧内编码而言, 除过保留了已有的宏块编码模式 I4MB、I 16MB和 I8MB之外,又新增添了两个宏块编码模式 I4MB_RD0T、 I 16MB—RD0T 和 I8MB_RD0T。对于新增的两个宏块编码模式, 编码器都需要对每个残差块选择 一个最佳的变换矩阵。
本发明实施例提供的视频编码方法, 可以根据帧内预测模式, 根据率失真 准则从多个候选变换矩阵中选择一组最优的变换矩阵对预测残差进行变换编 码, 得到变换结果。 通过这种方式进行编码, 可以根据每个残差块的特性, 针 对性地选择最有效的变换矩阵进行变换, 从而提高编码效率。
本发明实施例提供的视频数据编码器, 如图 4所示, 包括:
残差生成单元 401 , 用于根据输入的视频数据生成预测残差;
变换单元 402 , 用于根据帧内预测模式, 根据率失真准则从多个候选变换矩 阵中选择一组最优的变换矩阵对预测残差进行变换编码 , 得到变换结果;
码流生成单元 403, 用于才艮据变换结果和所选用的变换矩阵索引信息, 生成 编码码流。
其中, 变换单元 402 , 具体用于根据帧内预测模式, 遍历多个候选变换矩阵 中的所有列变换矩阵和行变换矩阵的组合, 选择矩阵相乘后率失真代价最 d、的 变换组合作为最优的变换系数矩阵, 并得到变换结果。
进一步地, 如图 5所示, 该视频数据编码器还包括:
系数扫描单元 501 ,具体用于根据帧内预测模式和所述变换矩阵索引信息选 择一组系数扫描顺序对变换后的系数进行扫描。
判断单元 502 ,用于确定以各种编码方式对预测残差进行编码后率失真代价 最小的模式作为帧内预测模式, 并得到编码结果。
索引编码单元 503, 用于将变换系数矩阵的索引信息写入编码数据中。
本发明实施例提供的视频编码器, 可以根据帧内预测模式, 根据率失真准 则从多个候选变换矩阵中选择一组最优的变换矩阵对预测残差进行变换编码 , 得到变换结果。 通过这种方式进行编码, 可以根据每个残差块的特性, 针对性 地选择最有效的变换矩阵进行变换, 从而提高编码效率。 下面结合图 4和图 5 ,对本发明实施例提供的视频解码方法进行进一步说明: 本发明实施例提供的视频数据编码器, 如图 4所示, 包括:
残差生成单元 401 , 用于根据输入的视频数据生成预测残差;
变换单元 402 , 用于根据帧内预测模式, 根据率失真准则从多个候选变换矩 阵中选择一组最优的变换矩阵对预测残差进行变换编码, 得到变换结果;
码流生成单元 403, 用于才艮据变换结果和所选用的变换矩阵索引信息, 生成 编码码流。
本发明实施例中, 选择的一组最优的变换矩阵可以为一个非分离变换矩阵; 也可以为一对变换矩阵, 即包括一个列变换矩阵和一个行变换矩阵。
本发明实施例中, 变换单元 402根据帧内预测模式, 根据率失真准则从多 个候选变换矩阵中选择一组最优的变换矩阵对预测残差进行变换编码, 得到变 换结果即为, 根据帧内预测模式, 利用多个候选变换矩阵对预测残差进行变换 编码, 根据率失真准则选择一组最优的变换矩阵, 并将这组最优变换矩阵对应 的变换结果用于后续和所选用的变换矩阵索引信息生成编码码流。
其中, 变换单元 402 , 具体用于根据帧内预测模式, 遍历多个候选变换矩阵 中的所有列变换矩阵和行变换矩阵的组合, 选择矩阵相乘后率失真代价最 、的 变换组合作为最优的变换矩阵, 并得到变换结果。 也即根据帧内预测模式, 遍 历多个候选变换矩阵中的所有列变换矩阵和行变换矩阵的组合, 选择残差变换 编码后率失真代价最 d、的变换组合作为最优的变换矩阵 , 并将这组最优变换矩 阵对应的变换结果用于后续和所选用的变换矩阵索引信息生成编码码流。
进一步地, 如图 5所示, 该视频数据编码器还包括:
系数扫描单元 501 ,具体用于根据帧内预测模式和所述变换矩阵索引信息选 择一组系数扫描顺序对变换后的系数进行扫描。
判断单元 502 ,用于确定以各种编码方式对预测残差进行编码后率失真代价 最小的模式作为帧内预测模式, 并得到编码结果。
索引编码单元 503, 用于将变换矩阵索引信息写入编码数据中。
若所述一组最优的变换矩阵为一对变换矩阵, 则将变换矩阵索引信息写入 编码数据中包括: 对一对变换矩阵的索引信息进行联合编码, 或对一对变换矩 阵的索引信息分别进行编码; 将索引信息编码结果写入编码数据中。 联合编码表明列变换矩阵和行变换矩阵成对出现 , 每一个行变换矩阵对应 限制, 例如一个行变换矩阵可以对应任意个一个列变换矩阵, 这样可以节省变 换矩阵的存储空间。
本发明实施例提供的视频编码器, 可以根据帧内预测模式, 根据率失真准 则从多个候选变换矩阵中选择一组最优的变换矩阵对预测残差进行变换编码 , 得到变换结果。 通过这种方式进行编码, 可以根据每个残差块的特性, 针对性 地选择最有效的变换矩阵进行变换 , 从而提高编码效率。
本发明实施例提供的视频解码器, 如图 6所示, 包括:
解析单元 601 , 用于对视频码流进行解析, 得到计算结果和编码变换系数矩 阵的索引信息;
确定单元 602 ,用于根据索引信息和帧内预测模式从多个候选变换矩阵中确 定变换系数矩阵;
重建单元 603, 用于利用变换系数矩阵对计算结果进行反变换, 得到残差数 据; 根据残差数据重建视频数据。
进一步地, 如图 7所述, 该视频解码器还包括:
反系数扫描单元 701 , 用于根据帧内预测模式和变换系数矩阵的索引信息, 选择一组系数扫描顺序对变换后的系数进行反系数扫描。 本发明实施例提供的视频解码器, 能够对视频编码码流进行解析, 得到计 算结果和编码变换系数矩阵的索引信息; 根据索引信息和所述帧内预测模式从 多个候选变换矩阵中确定变换系数矩阵 , 利用变换系数矩阵对计算结果进行反 变换, 得到残差数据, 根据所述残差数据重建视频数据。 这样, 可以再不增加 复杂度的情况下, 进行解码。 由于该编码釆用了上述实施例提供的方法, 能够 针对残差的特性选择最优的变换矩阵, 从而使熵编码效率有所提高, 再通过本 实施例提供的解码方法, 能够有效提高视频编解码的整体效率。
下面结合图 6和图 7 ,对本发明实施例提供的视频解码方法进行进一步说明: 本发明实施例提供的视频解码器, 如图 6所示, 包括:
解析单元 601 , 用于对视频码流进行解析, 得到计算结果和变换矩阵索引信 息;
确定单元 602 ,用于根据变换矩阵索引信息和帧内预测模式从多个候选变换 矩阵中确定变换矩阵;
重建单元 603 , 用于利用确定的变换矩阵对计算结果进行反变换, 得到残差 数据; 根据残差数据重建视频数据。
本发明实施例中, 解析单元 601 解析得到的结果包括变换结果, 也即本发 明实施例中用到的计算结果即为变换结果, 变换结果可以包括经过变换后得到 的变换系数矩阵。
本发明实施例中, 确定的变换矩阵是一组变换矩阵, 一组变换矩阵可以为 一个非分离变换矩阵; 也可以为一对变换矩阵, 即包括一个列变换矩阵和一个 行变换矩阵。
当编码变换过程中, 采用的是分离变换时, 确定单元 602 用于根据变换矩 阵索引信息中的行变换矩阵索引信息和列变换矩阵索引信息, 以及帧内预测模 式从一组候选的行变换矩阵和列变换矩阵中确定变换矩阵。 此处一组候选的行 变换矩阵和列变换矩阵可以包括多个行变换矩阵和列变换矩阵。 重建单元 603 利用行变换矩阵和列变换矩阵对计算结果进行反变换, 得到残差数据, 并根据 残差数据重建视频数据。
进一步地, 如图 7所述, 该视频解码器还包括:
反系数扫描单元 701 , 用于根据帧内预测模式和变换矩阵索引信息, 选择一 组系数扫描顺序对变换后的系数进行反系数扫描。
本发明实施例提供的视频解码器, 能够对视频编码码流进行解析, 得到计 算结果和变换矩阵索引信息; 根据索引信息和帧内预测模式从多个候选变换矩 阵中确定变换矩阵, 利用变换矩阵对计算结果进行反变换, 得到残差数据, 根 据所述残差数据重建视频数据。 这样, 可以在不增加复杂度的情况下, 进行解 码。 由于该编码采用了上述实施例提供的方法, 能够针对残差的特性选择最优 的变换矩阵, 从而使熵编码效率有所提高, 再通过本实施例提供的解码方法, 能够有效提高视频编解码的整体效率。
本发明实施例还提供一种视频数据编码方法, 如图 8 所示, 该方法步驟包 括:
S801、 根据输入的视频数据生成预测残差。
S802、 根据帧内预测模式, 根据最优化准则从多个候选变换矩阵中选择一 组最优的变换矩阵对预测残差进行变换编码, 得到变换结果。
本发明实施例中, 选择的一组最优的变换矩阵可以为一个非分离变换矩阵; 也可以为一对变换矩阵, 即包括一个列变换矩阵和一个行变换矩阵。
最优化准则可以为率失真准则、 绝对误差和 ( SAD , Sum of Abso lute Difference )、 编码比特或失真。 根据最优化准则选择可以包括多种方式, 例如 选择率失真代价最小的、 SAD最小的、 编码比特最少的或者是失真最小的等等。
本发明实施例中, 根据帧内预测模式, 根据最优化准则从多个候选变换矩 阵中选择一组最优的变换矩阵对预测残差进行变换编码 , 得到变换结果即为 , 根据帧内预测模式, 利用多个候选变换矩阵对预测残差进行变换编码, 根据最 优化准则选择一組最优的变换矩阵, 并将这組最优变换矩阵对应的变换结果用 于后续和所选用的变换矩阵索引信息生成编码码流。
某些实施方式中, 在进行变换的过程中, 还可以采用行列分离变换, 即根 据帧内预测模式, 遍历多个候选变换矩阵中的所有可能的列变换矩阵和行变换 矩阵的组合 , 选择残差变换编码后最优化准则代价最 、的变换组合作为最优变 换矩阵, 并得到变换结果。 也即根据帧内预测模式, 遍历多个候选变换矩阵中 的所有列变换矩阵和行变换矩阵的组合, 选择残差变换编码后最优化准则代价 最小的变换组合作为最优的变换矩阵, 并将这组最优变换矩阵对应的变换结果 用于后续和所选用的变换矩阵索引信息生成编码码流。
S803、 才艮据变换结果和所选用的变换矩阵索引信息, 生成编码码流。
进一步地, 本发明实施例还可以包括: 系数扫描过程, 为根据帧内预测模 式和变换矩阵索引信息选择一组系数扫描顺序对变换后的系数进行扫描。
然后选择变换后最优化准则代价最小的一种作为最佳帧内预测模式, 对其 结果量化后进行熵编码。 即以各种编码方式对所述预测残差进行编码, 以其中 最优化准则代价最小的模式作为帧内预测模式, 得到编码结果。
本发明实施例中 , 所述 居所述变换结果和所选用的变换矩阵索引信息 , 生成编码码流, 包括将变换矩阵索引信息写入编码数据中。
若所述一组最优的变换矩阵为一对变换矩阵, 则将变换矩阵索引信息写入 编码数据中包括: 对一对变换矩阵的索引信息进行联合编码, 或对一对变换矩 阵的索引信息分别进行编码; 将索引信息编码结果写入编码数据中。
限制, 例如一个行变换矩阵可以对应任意一个列变换矩阵, 这样可以节省变换 矩阵的存储空间。
本发明实施例提供的视频编码方法, 可以根据帧内预测模式, 根据最优化 准则从多个候选变换矩阵中选择一组最优的变换矩阵对预测残差进行变换编 码, 得到变换结果。 通过这种方式进行编码, 可以根据每个残差块的特性, 针 对性地选择最有效的变换矩阵进行变换, 从而提高编码效率。
本发明实施例还提供一种视频解码方法, 如图 9所示, 该方法包括:
S901、 对视频编码码流进行解析, 得到变换结果和变换矩阵索引信息。 进一步地, 本发明实施例还包括: 反系数扫描过程, 为根据帧内预测模式 和变换矩阵索引信息, 选择一组系数扫描顺序对该变换后的系数进行反系数扫 描。
S902、 根据变换矩阵索引信息和帧内预测模式从多个候选变换矩阵中确定 一组变换矩阵, 利用所述一组变换矩阵对变换结果进行反变换, 得到残差数据, 根据残差数据重建视频数据。
本发明实施例中, 确定的一组变换矩阵可以为一个非分离变换矩阵; 也可 以为一对变换矩阵, 即包括一个列变换矩阵和一个行变换矩阵。
本发明实施例中, 解析得到的结果包括变换结果, 也即本发明实施例中用 到的变换结果即为计算结果, 变换结果可以包括经过变换后得到的变换系数矩 阵。
具体的, 当编码变换过程中, 采用的是分离变换时, 本步骤 S902中的确定 的一组变换矩阵可以为根据变换矩阵索引信息中的行变换矩阵索引信息和列变 换矩阵索引信息, 以及帧内预测模式从多个候选的行变换矩阵和列变换矩阵中 确定的。
本发明实施例提供的视频解码方法, 能够对视频编码码流进行解析, 得到 变换结果和变换矩阵索引信息, 根据变换矩阵索引信息和帧内预测模式从多个 候选变换矩阵中确定一组变换矩阵, 利用变换矩阵对变换结果进行反变换, 得 到残差数据, 根据残差数据重建视频数据。 这样, 可以在不增加复杂度的情况 下, 进行解码。 由于编码釆用了上述实施例提供的方法, 能够针对残差的特性 选择最优的变换矩阵, 从而使熵编码效率有所提高, 再通过本实施例提供的解 码方法, 能够有效提高视频编解码的整体效率。
本发明实施例提供的视频数据编码方法, 如图 10所示, 该方法步驟包括: S100K 根据输入的视频数据生成预测残差。
S1002、 根据最优化准则从多个候选变换矩阵中选择一组最优的变换矩阵对 预测残差进行变换编码, 得到变换结果。
本发明实施例中, 选择的一組最优的变换矩阵可以为一个非分离变换矩阵; 也可以为一对变换矩阵, 即包括一个列变换矩阵和一个行变换矩阵。
最优化准则可以为率失真准则、 绝对误差和 ( SAD , Sum of Absolute Difference )、 编码比特或失真。 根据最优化准则选择可以包括多种方式, 例如 选择率失真代价最小的、 SAD最小的、 编码比特最少的或者是失真最小的等等。
本发明实施例中 , 根据最优化准则从多个候选变换矩阵中选择一组最优的 变换矩阵对预测残差进行变换编码, 得到变换结果即为, 利用多个候选变换矩 阵对预测残差进行变换编码, 根据最优化准则选择一组最优的变换矩阵, 并将 这组最优变换矩阵对应的变换结果用于后续和所选用的变换矩阵索引信息生成 编码码流。
某些实施方式中, 在进行变换的过程中, 还可以釆用行列分离变换, 即遍 历多个候选变换矩阵中的所有可能的列变换矩阵和行变换矩阵的组合, 选择残 差变换编码后最优化准则代价最小的变换组合作为最优变换矩阵, 并得到变换 结果。 也即遍历多个候选变换矩阵中的所有列变换矩阵和行变换矩阵的组合, 选择残差变换编码后最优化准则代价最小的变换组合作为最优的变换矩阵 , 并 将这组最优变换矩阵对应的变换结果用于后续和所选用的变换矩阵索引信息生 成编码码流。
S1003、 根据变换结果, 并根据帧内预测模式对所选用的变换矩阵索引信息 进行编码, 生成编码码;克。
进一步地, 本发明实施例还可以包括: 系数扫描过程, 为根据变换矩阵索 引信息选择一组系数扫描顺序对变换后的系数进行扫描。
然后选择变换后最优化准则代价最小的一种作为最佳帧内预测模式, 对其 结果量化后进行熵编码。 即以各种编码方式对所述预测残差进行编码, 以其中 最优化准则代价最小的模式作为帧内预测模式, 得到编码结果。
本发明实施例中, 根据所述变换结果,并根据帧内预测模式对所选用的变换 矩阵索引信息进行编码, 生成编码码流包括根据所选用的帧内预测模式, 选定 一种变换矩阵索引信息的编码方法, 将所述变换矩阵索引信息写入编码数据中。 对于不同的帧内预测模式, 可以采用不同的变换矩阵索引信息的编码方法, 将 变换矩阵索引信息写入编码数据中。 若所述一組最优的变换矩阵为一对变换矩 阵, 则将变换矩阵索引信息写入编码数据中包括: 对一对变换矩阵的索引信息 进行联合编码, 或对一对变换矩阵的索引信息分别进行编码; 根据帧内预测模 式将索引信息编码结果写入编码数据中。
限制, 例如一个行变换矩阵可以对应任意一个列变换矩阵, 这样可以节省变换 矩阵的存储空间。
本发明实施例提供的视频编码方法, 可以根据最优化准则从多个候选变换 矩阵中选择一组最优的变换矩阵对预测残差进行变换编码, 得到变换结果。 通 过这种方式进行编码, 可以根据每个残差块的特性, 针对性地选择最有效的变 换矩阵进行变换, 从而提高编码效率。
本发明实施例提供的视频解码方法, 如图 11所示, 该方法包括:
S110 对视频编码码流进行解析, 得到变换结果, 并根据帧内预测模式得 到变换矩阵索引信息。
本发明实施例中, 解析得到的结果包括变换结果, 也即本发明实施例中用 到的变换结果即为计算结果, 变换结果可以包括经过变换后得到的变换系数矩 阵。 根据帧内预测模式得到变换矩阵索引信息包括: 根据所述帧内预测模式, 选定一种变换矩阵索引信息的解码方法, 解码得到所述变换矩阵索引信息。 对 于不同的帧内预测模式, 可以采用不同的解析方法, 对视频码流进行解析得到 变换矩阵索引信息。
进一步地, 本发明实施例还包括: 反系数扫描过程, 为根据变换矩阵索引 信息, 选择一组系数扫描顺序对该变换后的系数进行反系数扫描。
S1102、 根据变换矩阵索引信息从多个候选变换矩阵中确定变换矩阵, 利用 确定的变换矩阵对变换结果进行反变换, 得到残差数据, 根据残差数据重建视 频数据。
本发明实施例中, 确定的变换矩阵可以是一组变换矩阵, 一组变换矩阵可 以为一个非分离变换矩阵; 也可以为一对变换矩阵, 即包括一个列变换矩阵和 一个行变换矩阵。
具体的, 当解码变换过程中, 采用的是分离变换时, 本步驟 S1102 中的变 换矩阵可以为根据索引信息中的行变换矩阵索引信息和列变换矩阵索引信息, 从一组候选的行变换矩阵和列变换矩阵中确定的。 此处一组候选的行变换矩阵 和列变换矩阵包括多个行变换矩阵和列变换矩阵。
本发明实施例提供的视频解码方法, 能够对视频编码码流进行解析, 得到 变换结果, 并根据帧内预测模式解析得到变换矩阵索引信息, 根据变换矩阵索 引信息从多个候选变换矩阵中确定变换矩阵, 利用变换矩阵对变换结果进行反 变换, 得到残差数据, 根据残差数据重建视频数据。 这样, 可以在不增加复杂 度的情况下, 进行解码。 由于该编码采用了上述实施例提供的方法, 能够针对 残差的特性选择最优的变换矩阵, 从而使熵编码效率有所提高, 再通过本实施 例提供的解码方法, 能够有效提高视频编解码的整体效率。
本发明实施例提供的视频数据编码器, 如图 12所示, 包括:
残差生成单元 1201 , 用于根据输入的视频数据生成预测残差; 变换单元 1202 , 根据最优化准则从多个候选变换矩阵中选择一组最优的变 换矩阵对预测残差进行变换编码, 得到变换结果;
码流生成单元 1203 , 用于根据所述变换结果, 并根据帧内预测模式对所选 用的变换矩阵索引信息进行编码, 生成编码码流。
本发明实施例中, 选择的一组最优的变换矩阵可以为一个非分离变换矩阵; 也可以为一对变换矩阵, 即包括一个列变换矩阵和一个行变换矩阵。 所述的最 优化准则包括: 率失真准则、 绝对误差和 SAD、 编码比特或失真
本发明实施例中, 变换单元 1202具体用于遍历多个候选变换矩阵中的所有 列变换矩阵和行变换矩阵的组合, 选择残差变换编码后最优化准则代价最小的 变换组合作为最优的变换矩阵, 并得到变换结果。
本发明实施例中, 码流生成单元 1203才艮据所述变换结果, 并才艮据帧内预测 模式对所选用的变换矩阵索引信息进行编码, 生成编码码流包括: 根据所选用 的帧内预测模式, 选定一种变换矩阵索引信息的编码方法, 将所述变换矩阵索 引信息写入编码数据中。
进一步地, 如图 13所示, 该视频数据编码器还包括:
系数扫描单元 1301 , 用于根据所述变换矩阵索引信息选择一组系数扫描顺 序对变换后的系数进行扫描。
判断单元 1302 , 用于确定以各种编码方式对所述预测残差进行编码后最优 化准则代价最小的模式作为帧内预测模式, 并得到编码结果。
索引编码单元 1303 , 用于根据所选用的帧内预测模式, 选定一种变换矩阵 索引信息的编码方法, 将所述变换矩阵索引信息写入编码数据中。
若所述一组最优的变换矩阵为一对变换矩阵, 则根据所选用的帧内预测模 式, 选定一种变换矩阵索引信息的编码方法, 将所述变换矩阵索引信息写入编 码数据中包括: 对一对变换矩阵的索引信息进行联合编码, 或对一对变换矩阵 的索引信息分别进行编码; 根据所选用的帧内预测模式, 选定一种变换矩阵索 引信息的编码方法, 将所述变换矩阵索引信息写入编码数据中。 一个相应的列变换矩阵; 分别编码表明列变换矩阵和行变换矩阵没有对应性的 限制, 例如一个行变换矩阵可以对应任意个一个列变换矩阵, 这样可以节省变 换矩阵的存储空间。
本发明实施例提供的视频编码器, 可以根据最优化准则从多个候选变换矩 阵中选择一组最优的变换矩阵对预测残差进行变换编码, 得到变换结果。 通过 这种方式进行编码, 可以根据每个残差块的特性, 针对性地选择最有效的变换 矩阵进行变换, 从而提高编码效率。
本发明实施例提供的视频编码器, 可以根据帧内预测模式, 根据率失真准 则从多个候选变换矩阵中选择一组最优的变换矩阵对预测残差进行变换编码, 得到变换结果。 通过这种方式进行编码, 可以根据每个残差块的特性, 针对性 地选择最有效的变换矩阵进行变换, 从而提高编码效率。
本发明实施例提供的视频解码器, 如图 14所示, 包括:
解析单元 1401 , 用于对视频码流进行解析, 得到变换结果, 并根据帧内预 测模式得到变换矩阵索引信息;
确定单元 1402 , 用于根据所述变换矩阵索引信息从多个候选变换矩阵中确 定变换矩阵;
重建单元 1403 , 用于利用确定的变换矩阵对所述变换结果进行反变换, 得 到残差数据; 根据所述残差数据重建视频数据。
本发明实施例中, 确定的变换矩阵是一组变换矩阵, 一组变换矩阵可以为 一个非分离变换矩阵; 也可以为一对变换矩阵, 即包括一个列变换矩阵和一个 行变换矩阵。
本发明实施例中, 解析得到的结果包括变换结果, 也即本发明实施例中用 到的变换结果即为计算结果, 变换结果可以包括经过变换后得到的变换系数矩 阵。
解析单元 1401根据帧内预测模式得到变换矩阵索引信息包括: 根据所述帧 内预测模式, 选定一种变换矩阵索引信息的解码方法, 解码得到所述变换矩阵 索引信息。
进一步地, 如图 15所述, 该视频解码器还包括:
反系数扫描单元 1501 , 用于根据所述变换矩阵索引信息选择一组系数扫描 顺序对变换后的系数进行反系数扫描。
本发明实施例提供的视频解码器, 能够对视频编码码流进行解析, 得到变 换结果, 并根据帧内预测模式解析得到变换矩阵索引信息, 根据变换矩阵索引 信息从多个候选变换矩阵中确定变换矩阵, 利用变换矩阵对变换结果进行反变 换, 得到残差数据, 根据残差数据重建视频数据。 这样, 可以在不增加复杂度 的情况下, 进行解码。 由于该编码采用了上述实施例提供的方法, 能够针对残 差的特性选择最优的变换矩阵, 从而使熵编码效率有所提高, 再通过本实施例 提供的解码方法, 能够有效提高视频编解码的整体效率。
本领域普通技术人员可以理解: 实现上述方法实施例的全部或部分步驟可 以通过程序指令相关的硬件来完成, 前述的程序可以存储于一计算机可读取存 储介质中, 该程序在执行时, 执行包括上述方法实施例的步驟, 而前述的存储 介质包括: ROM、 RAM, 磁碟或者光盘等各种可以存储程序代码的介质。
以上所述, 仅为本发明的具体实施方式, 但本发明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发明揭露的技术范围内, 可轻易想到 变化或替换, 都应涵盖在本发明的保护范围之内。 因此, 本发明的保护范围应 所述以权利要求的保护范围为准。

Claims

权 利 要 求 书
1、 一种视频数据编码方法, 其特征在于, 包括:
根据输入的视频数据生成预测残差;
根据帧内预测模式, 根据率失真准则从多个候选变换矩阵中选择一组最优 的变换矩阵对预测残差进行变换编码, 得到变换结果;
才艮据所述变换结果和所选用的变换矩阵索引信息, 生成编码码流。
2、 根据权利要求 1所述的视频数据编码方法, 其特征在于, 所述方法还包 括:
根据所述帧内预测模式和所述变换矩阵索引信息选择一组系数扫描顺序对 变换后的系数进行扫描。
3、 根据权利要求 1所述的视频数据编码方法, 其特征在于, 所述方法还包 括:
以各种编码方式对所述预测残差进行编码 , 以其中率失真代价最小的模式 作为帧内预测模式, 得到编码结果。
4、 根据权利要求 1所述的视频数据编码方法, 其特征在于, 所述根据所述 变换编码后的预测残差和所选用的变换矩阵索引信息, 生成编码码流, 包括: 将所述变换系数矩阵的索引信息写入编码数据中。
5、 根据权利要求 1所述的视频数据编码方法, 其特征在于, 所述根据帧内 预测模式, 根据率失真准则从多个候选变换矩阵中选择一组最优的变换矩阵对 预测残差进行变换编码, 得到变换结果, 包括:
根据帧内预测模式, 遍历多个候选变换矩阵中的所有列变换矩阵和行变换 矩阵的组合, 选择矩阵相乘后率失真代价最小的变换组合作为最优的变换矩阵, 并得到变换结果。
6、 根据权利要求 1所述的视频数据编码方法, 其特征在于:
所述一组最优的变换矩阵为一个非分离变换矩阵; 或者,
所述一组最优的变换矩阵为一对变换矩阵, 所述一对变换矩阵包括一个列 变换矩阵和一个行变换矩阵。
7、 根据权利要求 6所述的视频数据编码方法, 其特征在于, 所述根据所述 变换结果和所选用的变换矩阵索引信息, 生成编码码流, 包括:
将所述变换矩阵索引信息写入编码数据中。
8、 根据权利要求 4或 7所述的视频数据编码方法, 其特征在于, 若所述一 组最优的变换矩阵为一对变换矩阵, 所述将所述变换矩阵索引信息写入编码数 据中包括:
对一对变换矩阵的索引信息进行联合编码, 或对一对变换矩阵的索引信息 分别进行编码;
将索引信息编码结果写入编码数据中。
9、 一种视频解码方法, 其特征在于, 包括:
对视频编码码流进行解析, 得到计算结果和编码变换系数矩阵的索引信息; 根据所述索引信息和帧内预测模式从多个候选变换矩阵中确定变换系数矩 阵, 利用所述变换系数矩阵对所述计算结果进行反变换, 得到残差数据, 根据 所述残差数据重建视频数据。
10、 根据权利要求 9所述的视频解码方法, 其特征在于, 所述方法还包括: 根据所述帧内预测模式和所述变换系数矩阵的索引信息, 选择一組系数扫描顺 序对变换后的系数进行反系数扫描。
11、 根据权利要求 9 所述的视频解码方法, 其特征在于, 所述变换系数矩 阵为根据所述索引信息中的行变换系数矩阵索引信息和列变换系数矩阵索引信 息, 以及所述帧内预测模式从一组候选的行变换矩阵和列变换矩阵中确定的。
12、 根据权利要求 9所述的视频解码方法, 其特征在于:
所述确定的变换矩阵为一个非分离变换矩阵; 或者,
所述确定的变换矩阵为一对变换矩阵, 所述一对变换矩阵包括一个列变换 矩阵和一个行变换矩阵。
13、 根据权利要求 12所述的视频解码方法, 其特征在于, 变换矩阵为根据 所述变换矩阵索引信息中的行变换矩阵索引信息和列变换矩阵索引信息, 以及 所述帧内预测模式从一组候选的行变换矩阵和列变换矩阵中确定的。
14、 一种视频数据编码器, 其特征在于, 包括:
残差生成单元, 用于根据输入的视频数据生成预测残差; 变换单元, 用于根据帧内预测模式, 根据率失真准则从多个候选变换矩阵 中选择一组最优的变换矩阵对预测残差进行变换编码, 得到变换结果;
码流生成单元, 用于^据所述变换结果和所选用的变换矩阵索引信息, 生 成编码码流。
15、 根据权利要求 14所述的视频数据编码器, 其特征在于, 所述视频数据 编码器还包括:
系数扫描单元, 具体用于根据所述帧内预测模式和所述变换矩阵索引信息 选择一组系数扫描顺序对变换后的系数进行扫描。
16、 根据权利要求 14所述的视频数据编码器, 其特征在于, 所述视频数据 编码器还包括:
判断单元, 用于确定以各种编码方式对所述预测残差进行编码后率失真代 价最小的模式作为帧内预测模式, 并得到编码结果。
17、 根据权利要求 14所述的视频数据编码器, 其特征在于, 所述视频数据 编码器还包括:
索引编码单元, 用于将所述变换系数矩阵的索引信息写入编码数据中。
18、根据权利要求 14所述的视频数据编码器,其特征在于, 所述变换单元, 具体用于根据帧内预测模式, 遍历多个候选变换矩阵中的所有列变换矩阵和行 变换矩阵的组合, 选择矩阵相乘后率失真代价最 d、的变换组合作为最优的变换 矩阵, 并得到变换结果。
19、 一种视频解码器, 其特征在于, 包括:
解析单元, 用于对视频码流进行解析, 得到计算结果和编码变换系数矩阵 的索引信息;
确定单元, 用于根据所述索引信息和所述帧内预测模式从多个候选变换矩 阵中确定变换系数矩阵;
重建单元, 用于利用所述变换系数矩阵对所述计算结果进行反变换, 得到 残差数据; 根据所述残差数据重建视频数据。
20、 根据权利要求 19所述的视频解码器, 其特征在于, 所述视频解码器还 包括: 反系数扫描单元, 用于根据所述帧内预测模式和所述变换系数矩阵的索 引信息, 选择一组系数扫描顺序对变换后的系数进行反系数扫描。
21、 根据权利要求 19所述的视频解码器, 其特征在于, 所述确定单元具体 用于根据所述索引信息中的行变换系数矩阵索引信息和列变换系数矩阵索引信 息, 以及所述帧内预测模式从一组候选的行变换矩阵和列变换矩阵中确定变换 矩阵。
22、 一种视频数据编码方法, 其特征在于, 包括:
根据输入的视频数据生成预测残差;
根据帧内预测模式, 根据最优化准则从多个候选变换矩阵中选择一组最优 的变换矩阵对预测残差进行变换编码 , 得到变换结果;
才艮据所述变换结果和所选用的变换矩阵索引信息, 生成编码码流。
23、 根据权利要求 22所述的视频数据编码方法, 其特征在于:
所述一组最优的变换矩阵为一个非分离变换矩阵; 或者,
所述一组最优的变换矩阵为一对变换矩阵, 所述一对变换矩阵包括一个列 变换矩阵和一个行变换矩阵。
24、 根据权利要求 22所述的视频数据编码方法, 其特征在于, 所述的最优 化准则包括: 率失真准则、 绝对误差和 SAD、 编码比特或失真。
25、 根据权利要求 22所述的视频数据编码方法, 其特征在于, 所述方法还 包括:
根据所述帧内预测模式和所述变换矩阵索引信息选择一组系数扫描顺序对 变换后的系数进行扫描。
26、 根据权利要求 22所述的视频数据编码方法, 其特征在于, 所述方法还 包括:
以各种编码方式对所述预测残差进行编码 , 以其中最优化准则代价最小的 模式作为帧内预测模式, 得到编码结果。
27、 根据权利要求 22或 23所述的视频数据编码方法, 其特征在于, 所述 根据所述变换结果和所选用的变换矩阵索引信息, 生成编码码流, 包括:
将所述变换矩阵索引信息写入编码数据中。
28、 根据权利要求 27所述的视频数据编码方法, 其特征在于, 若所述一组 最优的变换矩阵为一对变换矩阵 , 所述将所述变换矩阵索引信息写入编码数据 中包括:
对一对变换矩阵的索引信息进行联合编码, 或对一对变换矩阵的索引信息 分别进行编码;
将索引信息编码结果写入编码数据中。
29、 根据权利要求 22所述的视频数据编码方法, 其特征在于, 所述根据帧 内预测模式, 根据最优化准则从多个候选变换矩阵中选择一组最优的变换矩阵 对预测残差进行变换编码, 得到变换结果, 包括:
根据帧内预测模式, 遍历多个候选变换矩阵中的所有列变换矩阵和行变换 矩阵的组合, 选择残差变换编码后最优化准则代价最小的变换组合作为最优的 变换矩阵, 并得到变换结果。
30、 一种视频解码方法, 其特征在于, 包括:
对视频编码码流进行解析, 得到变换结果和变换矩阵索引信息;
根据所述变换矩阵索引信息和帧内预测模式从多个候选变换矩阵中确定一 組变换矩阵, 利用所述一组变换矩阵对所述变换结果进行反变换, 得到残差数 据, 根据所述残差数据重建视频数据。
31、 根据权利要求 30所述的视频解码方法, 其特征在于:
所述一组变换矩阵为一个非分离变换矩阵; 或者,
所述一组变换矩阵为一对变换矩阵, 所述一对变换矩阵包括一个列变换矩 阵和一个行变换矩阵。
32、根据权利要求 30所述的视频解码方法,其特征在于, 所述方法还包括: 根据所述帧内预测模式和所述变换矩阵索引信息, 选择一组系数扫描顺序对变 换后的系数进行反系数扫描。
33、 根据权利要求 30或 31所述的视频解码方法, 其特征在于, 所述一组 变换矩阵为根据所述变换矩阵索引信息中的行变换矩阵索引信息和列变换矩阵 索引信息, 以及所述帧内预测模式从多个候选的行变换矩阵和列变换矩阵中确 定的。
34、 一种视频数据编码方法, 其特征在于, 包括: 根据输入的视频数据生成预测残差;
根据最优化准则从多个候选变换矩阵中选择一组最优的变换矩阵对预测残 差进行变换编码, 得到变换结果;
根据所述变换结果,并根据帧内预测模式对所选用的变换矩阵索引信息进 行编码, 生成编码码流。
35、 根据权利要求 34所述的视频数据编码方法, 其特征在于:
所述一组最优的变换矩阵为一个非分离变换矩阵; 或者,
所述一组最优的变换矩阵为一对变换矩阵, 所述一对变换矩阵包括一个列 变换矩阵和一个行变换矩阵。
36、 根据权利要求 34所述的视频数据编码方法, 其特征在于, 所述的最优 化准则包括: 率失真准则、 绝对误差和 SAD、 编码比特或失真。
37、 根据权利要求 34所述的视频数据编码方法, 其特征在于, 所述方法还 包括:
根据所述变换矩阵索引信息选择一组系数扫描顺序对变换后的系数进行扫 描。
38、 根据权利要求 34所述的视频数据编码方法, 其特征在于, 所述方法还 包括:
以各种编码方式对所述预测残差进行编码, 以其中最优化准则代价最小的 模式作为帧内预测模式, 得到编码结果。
39、 根据权利要求 34所述的视频数据编码方法, 其特征在于, 所述根据所 述变换结果,并根据帧内预测模式对所选用的变换矩阵索引信息进行编码, 生成 编码码流, 包括:
根据所选用的帧内预测模式, 选定一种变换矩阵索引信息的编码方法, 将 所述变换矩阵索引信息写入编码数据中。
40、 根据权利要求 39所述的视频数据编码方法, 其特征在于, 若所述一组 最优的变换矩阵为一对变换矩阵, 所述根据所选用的帧内预测模式, 选定一种 变换矩阵索引信息的编码方法, 将所述变换矩阵索引信息写入编码数据中包括: 对一对变换矩阵的索引信息进行联合编码, 或对一对变换矩阵的索引信息 分别进行编码;
根据所选用的帧内预测模式, 选定一种变换矩阵索引信息的编码方法, 将 所述变换矩阵索引信息写入编码数据中。
41、 根据权利要求 34所述的视频数据编码方法, 其特征在于, 所述根据最 优化准则从多个候选变换矩阵中选择一组最优的变换矩阵对预测残差进行变换 编码, 得到变换结果, 包括:
差变换编码后最优化准则代价最小的变换组合作为最优的变换矩阵, 并得到变 换结果。
42、 一种视频解码方法, 其特征在于, 包括:
对视频编码码流进行解析, 得到变换结果, 并根据帧内预测模式得到变换 矩阵索引信息;
根据所述变换矩阵索引信息从多个候选变换矩阵中确定变换矩阵 , 利用确 定的变换矩阵对所述变换结果进行反变换, 得到残差数据, 根据所述残差数据 重建视频数据。
43、 根据权利要求 42所述的视频解码方法, 其特征在于:
所述确定的变换矩阵为一个非分离变换矩阵; 或者,
所述确定的变换矩阵为一对变换矩阵, 所述一对变换矩阵包括一个列变换 矩阵和一个行变换矩阵。
44、 根据权利要求 42所述的视频解码方法, 其特征在于, 所述根据帧内预 测模式得到变换矩阵索引信息包括:
根据所述帧内预测模式, 选定一种变换矩阵索引信息的解码方法, 解码得 到所述变换矩阵索引信息。
45、根据权利要求 42所述的视频解码方法,其特征在于, 所述方法还包括: 根据变换矩阵索引信息选择一组系数扫描顺序对变换后的系数进行反系数扫 描。
46、 根据权利要求 42或 43所述的视频解码方法, 其特征在于, 变换矩阵 为根据所述变换矩阵索引信息中的行变换矩阵索引信息和列变换矩阵索引信 息, 从一组候选的行变换矩阵和列变换矩阵中确定的。
47、 一种视频数据编码器, 其特征在于, 包括:
残差生成单元, 用于根据输入的视频数据生成预测残差;
变换单元, 用于根据最优化准则从多个候选变换矩阵中选择一组最优的变 换矩阵对预测残差进行变换编码, 得到变换结果;
码流生成单元, 用于根据所述变换结果, 并根据帧内预测模式对所选用的 变换矩阵索引信息进行编码, 生成编码码流。
48、 根据权利要求 47所述的视频数据编码器, 其特征在于, 所述视频数据 编码器还包括:
系数扫描单元, 用于根据所述变换矩阵索引信息选择一组系数扫描顺序对 变换后的系数进行扫描。
49、 根据权利要求 47所述的视频数据编码器, 其特征在于, 所述视频数据 编码器还包括:
判断单元, 用于确定以各种编码方式对所述预测残差进行编码后最优化准 则代价最小的模式作为帧内预测模式, 并得到编码结果。
50、 根据权利要求 47所述的视频数据编码器, 其特征在于, 所述视频数据 编码器还包括:
索引编码单元, 用于根据所选用的帧内预测模式, 选定一种变换矩阵索引 信息的编码方法, 将所述变换矩阵索引信息写入编码数据中。
51、根据权利要求 47所述的视频数据编码器,其特征在于, 所述变换单元, 具体用于遍历多个候选变换矩阵中的所有列变换矩阵和行变换矩阵的组合, 选 择残差变换编码后最优化准则代价最小的变换组合作为最优的变换矩阵 , 并得 到变换结果。
52、 一种视频解码器, 其特征在于, 包括:
解析单元, 用于对视频码流进行解析, 得到变换结果, 并根据帧内预测模 式得到变换矩阵索引信息;
确定单元 , 用于根据所述变换矩阵索引信息从多个候选变换矩阵中确定变 换矩阵;
重建单元, 用于利用确定的变换矩阵对所述变换结果进行反变换, 得到残 差数据; 根据所述残差数据重建视频数据。
53、 根据权利要求 52所述的视频解码器, 其特征在于, 所述视频解码器还 包括: 反系数扫描单元, 用于根据所述变换矩阵索引信息选择一组系数扫描顺 序对变换后的系数进行反系数扫描。
54、 根据权利要求 52所述的视频解码器, 其特征在于, 所述确定单元具体 用于根据所述变换矩阵索引信息中的行变换矩阵索引信息和列变换矩阵索引信 息, 从一组候选的行变换矩阵和列变换矩阵中确定变换矩阵。
PCT/CN2010/076464 2009-10-23 2010-08-30 一种视频编解码方法及设备 WO2011047579A1 (zh)

Priority Applications (5)

Application Number Priority Date Filing Date Title
KR1020127011936A KR101481642B1 (ko) 2009-10-23 2010-08-30 비디오를 인코딩하고 디코딩하기 위한 방법 및 장치
EP10824420.3A EP2493197A4 (en) 2009-10-23 2010-08-30 METHOD AND DEVICE FOR VIDEO ENCODING AND DECODING
BR112012011325-9A BR112012011325B1 (pt) 2009-10-23 2010-08-30 Método para codificar dados de vídeo, método para decodificar vídeo, codificador de dados de vídeo e decodificador de vídeo
AU2010310286A AU2010310286B2 (en) 2009-10-23 2010-08-30 Method and device for encoding and decoding video
US13/452,198 US9723313B2 (en) 2009-10-23 2012-04-20 Method and device for encoding and decoding videos using a best transformation matrix among multiple transformation matrices

Applications Claiming Priority (6)

Application Number Priority Date Filing Date Title
CN200910209013.9 2009-10-23
CN200910209013 2009-10-23
CN201010147581 2010-04-09
CN201010147581.3 2010-04-09
CN2010102137918A CN102045560B (zh) 2009-10-23 2010-06-17 一种视频编解码方法及设备
CN201010213791.8 2010-06-17

Related Child Applications (1)

Application Number Title Priority Date Filing Date
US13/452,198 Continuation US9723313B2 (en) 2009-10-23 2012-04-20 Method and device for encoding and decoding videos using a best transformation matrix among multiple transformation matrices

Publications (1)

Publication Number Publication Date
WO2011047579A1 true WO2011047579A1 (zh) 2011-04-28

Family

ID=43899810

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2010/076464 WO2011047579A1 (zh) 2009-10-23 2010-08-30 一种视频编解码方法及设备

Country Status (7)

Country Link
US (1) US9723313B2 (zh)
EP (1) EP2493197A4 (zh)
KR (1) KR101481642B1 (zh)
CN (1) CN102045560B (zh)
AU (1) AU2010310286B2 (zh)
BR (1) BR112012011325B1 (zh)
WO (1) WO2011047579A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113273214A (zh) * 2018-12-19 2021-08-17 Lg电子株式会社 基于二次变换的图像编码方法及其装置

Families Citing this family (54)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102045560B (zh) 2009-10-23 2013-08-07 华为技术有限公司 一种视频编解码方法及设备
US8792740B2 (en) * 2010-02-02 2014-07-29 Humax Holdings Co., Ltd. Image encoding/decoding method for rate-distortion optimization and apparatus for performing same
CN102281435B (zh) * 2010-06-11 2013-10-02 华为技术有限公司 编码方法、解码方法、编码装置、解码装置及编解码系统
US20120163456A1 (en) 2010-12-22 2012-06-28 Qualcomm Incorporated Using a most probable scanning order to efficiently code scanning order information for a video block in video coding
US9049444B2 (en) * 2010-12-22 2015-06-02 Qualcomm Incorporated Mode dependent scanning of coefficients of a block of video data
ES2657197T3 (es) 2011-06-28 2018-03-01 Samsung Electronics Co., Ltd. Aparato de decodificación de video con intra predicción
CN102857755B (zh) * 2011-07-01 2016-12-14 华为技术有限公司 确定变换块尺寸的方法和设备
KR101641863B1 (ko) 2011-10-19 2016-07-22 주식회사 케이티 영상 부호화/복호화 방법 및 그 장치
KR20130049522A (ko) * 2011-11-04 2013-05-14 오수미 인트라 예측 블록 생성 방법
CN103096053B (zh) 2011-11-04 2015-10-07 华为技术有限公司 一种变换模式的编解码方法和装置
WO2013109066A1 (ko) * 2012-01-20 2013-07-25 주식회사 팬택 화면 내 예측 모드 매핑 방법 및 이러한 방법을 사용하는 장치
CN103533324B (zh) * 2012-07-03 2017-04-05 乐金电子(中国)研究开发中心有限公司 一种深度图像帧内编码方法、装置及编码器
KR101431463B1 (ko) * 2012-07-11 2014-08-22 세종대학교산학협력단 무손실 비디오 부호화/복호화 방법 및 장치
US10230956B2 (en) 2012-09-26 2019-03-12 Integrated Device Technology, Inc. Apparatuses and methods for optimizing rate-distortion of syntax elements
CN108200439B (zh) * 2013-06-14 2020-08-21 浙江大学 提高数字信号变换性能的方法及数字信号变换方法和装置
CN104853196B (zh) * 2014-02-18 2018-10-19 华为技术有限公司 编解码方法和装置
US20170280140A1 (en) * 2014-09-19 2017-09-28 Lg Electronics Inc. Method and apparatus for adaptively encoding, decoding a video signal based on separable transform
US20180115787A1 (en) * 2015-04-12 2018-04-26 Lg Electronics Inc. Method for encoding and decoding video signal, and apparatus therefor
FR3035761A1 (fr) * 2015-04-30 2016-11-04 Orange Procede de codage et de decodage d'images, dispositif de codage et de decodage d'images et programmes d'ordinateur correspondants
FR3038196A1 (fr) * 2015-06-29 2016-12-30 B<>Com Procede de codage d'une image numerique, procede de decodage, dispositifs et programmes d'ordinateurs associes
EP3334163A4 (en) * 2015-08-06 2019-04-17 LG Electronics Inc. DEVICE AND METHOD FOR PERFORMING TRANSFORMATION USING SINGLETON COEFFICIENT UPDATE
CN108353193B (zh) * 2015-08-19 2022-07-15 Lg 电子株式会社 基于多个基于图的模型处理视频数据的方法和设备
EP4106333A1 (en) * 2016-02-12 2022-12-21 Samsung Electronics Co., Ltd. Image encoding method and apparatus, and image decoding method and apparatus
US10390048B2 (en) * 2016-02-15 2019-08-20 Qualcomm Incorporated Efficient transform coding using optimized compact multi-pass transforms
CN105791867B (zh) * 2016-03-23 2019-02-22 北京大学 基于边界自适应变换的优化视频数据编码方法
EP3485637A1 (en) * 2016-07-14 2019-05-22 Fraunhofer Gesellschaft zur Förderung der Angewand Predictive picture coding using transform-based residual coding
CN107920247A (zh) * 2016-10-07 2018-04-17 财团法人工业技术研究院 选择画面内预测模式的方法、视频编码装置及处理设备
EP3586511B1 (en) * 2017-03-16 2022-01-05 MediaTek Inc. Method and apparatus of enhanced multiple transforms and non-separable secondary transform for video coding
US10574959B2 (en) * 2017-07-05 2020-02-25 Qualcomm Incorporated Color remapping for non-4:4:4 format video content
CN109922348B (zh) * 2017-12-13 2020-09-18 华为技术有限公司 图像编解码方法和装置
WO2019194420A1 (ko) * 2018-04-01 2019-10-10 엘지전자 주식회사 변환 인디케이터에 기반한 영상 코딩 방법 및 그 장치
KR102636267B1 (ko) 2018-08-16 2024-02-14 베이징 바이트댄스 네트워크 테크놀로지 컴퍼니, 리미티드 변형 행렬 선택의 계수에 따른 코딩
CN114928745B (zh) * 2018-09-02 2024-04-19 Lg电子株式会社 信号编解码方法、计算机可读存储介质和数据传输方法
CN111758261B (zh) * 2018-09-02 2022-06-10 Lg电子株式会社 用于处理图像信号的方法和设备
CN110944177B (zh) * 2018-09-21 2024-03-01 华为技术有限公司 视频解码方法及视频解码器,视频编码方法及视频编码器
CN111225206B (zh) * 2018-11-23 2021-10-26 华为技术有限公司 视频解码方法和视频解码器
CN111277840B (zh) * 2018-12-04 2022-02-08 华为技术有限公司 变换方法、反变换方法以及视频编码器和视频解码器
KR20210098967A (ko) 2019-01-01 2021-08-11 엘지전자 주식회사 이차 변환에 기반한 영상 코딩 방법 및 그 장치
CN109819250B (zh) * 2019-01-15 2020-09-25 北京大学 一种多核全组合方式的变换方法和系统
CN109788286B (zh) * 2019-02-01 2021-06-18 北京大学深圳研究生院 一种编码、解码变换方法、系统、设备及计算机可读介质
KR20220084194A (ko) 2019-03-26 2022-06-21 엘지전자 주식회사 변환에 기반한 영상 코딩 방법 및 그 장치
EP3939269A4 (en) * 2019-04-10 2022-06-15 Beijing Dajia Internet Information Technology Co., Ltd. METHOD AND APPARATUS FOR VIDEO CODING USING AN IMPROVED MATRIX-BASED INTRA PREDICTION CODING MODE
SG11202110936PA (en) 2019-04-12 2021-11-29 Beijing Bytedance Network Technology Co Ltd Chroma coding mode determination based on matrix-based intra prediction
JP7403555B2 (ja) 2019-04-16 2023-12-22 北京字節跳動網絡技術有限公司 イントラコーディングモードにおけるマトリクスの導出
CN113812150B (zh) 2019-05-01 2023-11-28 北京字节跳动网络技术有限公司 使用滤波的基于矩阵的帧内预测
CN117097912A (zh) 2019-05-01 2023-11-21 北京字节跳动网络技术有限公司 基于矩阵的帧内预测的上下文编码
BR112021022868A2 (pt) 2019-05-22 2022-01-04 Beijing Bytedance Network Tech Co Ltd Método de processamento de vídeos, aparelho para processar dados de vídeo e meios de armazenamento e gravação não transitórios legíveis por computador
CN113924775B (zh) 2019-05-31 2023-11-14 北京字节跳动网络技术有限公司 基于矩阵的帧内预测中的限制的上采样
WO2020244610A1 (en) 2019-06-05 2020-12-10 Beijing Bytedance Network Technology Co., Ltd. Context determination for matrix-based intra prediction
EP4032274A4 (en) * 2019-09-19 2023-11-01 Telefonaktiebolaget Lm Ericsson (Publ) METHOD ALLOWING AN INTRA PREDICTION BLOCK BASED ON A MATRIX TO COMPRISE MULTIPLE TRANSFORMATION BLOCKS
EP4042689A4 (en) 2019-10-28 2023-06-07 Beijing Bytedance Network Technology Co., Ltd. SIGNALING AND SYNTAX ANALYSIS BASED ON A COLOR COMPONENT
CN112565751B (zh) * 2020-09-27 2021-09-10 腾讯科技(深圳)有限公司 视频解码方法、装置、计算机可读介质及电子设备
US20230078100A1 (en) * 2021-08-30 2023-03-16 Tencent America LLC Scan order of secondary transform coefficients
CN117831545A (zh) * 2022-09-29 2024-04-05 抖音视界有限公司 编码、解码方法、编码器、解码器、电子设备和存储介质

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1564602A (zh) * 2004-03-18 2005-01-12 华中科技大学 视频编码的整数变换矩阵选择方法及相关的整数变换方法
US20060133509A1 (en) * 2004-12-16 2006-06-22 Schwartz Mayer D Methods of selecting an encoding mode
WO2008132890A1 (ja) * 2007-04-16 2008-11-06 Kabushiki Kaisha Toshiba 画像符号化と画像復号化の方法及び装置
CN101489134A (zh) * 2009-01-16 2009-07-22 华中科技大学 用于视频帧内编码的klt矩阵训练方法

Family Cites Families (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP3887178B2 (ja) 2001-04-09 2007-02-28 株式会社エヌ・ティ・ティ・ドコモ 信号符号化方法及び装置並びに復号方法及び装置
JP4447197B2 (ja) 2002-01-07 2010-04-07 三菱電機株式会社 動画像符号化装置および動画像復号装置
US20050213835A1 (en) * 2004-03-18 2005-09-29 Huazhong University Of Science & Technology And Samsung Electronics Co., Ltd. Integer transform matrix selection method in video coding and related integer transform method
WO2006028088A1 (ja) * 2004-09-08 2006-03-16 Matsushita Electric Industrial Co., Ltd. 動画像符号化方法および動画像復号化方法
CN100564602C (zh) * 2006-07-05 2009-12-02 中国石油化工股份有限公司 一种氯铑酸的制备方法
US8488672B2 (en) 2007-04-17 2013-07-16 Qualcomm Incorporated Mode uniformity signaling for intra-coding
US8428133B2 (en) * 2007-06-15 2013-04-23 Qualcomm Incorporated Adaptive coding of video block prediction mode
CN102045560B (zh) 2009-10-23 2013-08-07 华为技术有限公司 一种视频编解码方法及设备

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1564602A (zh) * 2004-03-18 2005-01-12 华中科技大学 视频编码的整数变换矩阵选择方法及相关的整数变换方法
US20060133509A1 (en) * 2004-12-16 2006-06-22 Schwartz Mayer D Methods of selecting an encoding mode
WO2008132890A1 (ja) * 2007-04-16 2008-11-06 Kabushiki Kaisha Toshiba 画像符号化と画像復号化の方法及び装置
CN101489134A (zh) * 2009-01-16 2009-07-22 华中科技大学 用于视频帧内编码的klt矩阵训练方法

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113273214A (zh) * 2018-12-19 2021-08-17 Lg电子株式会社 基于二次变换的图像编码方法及其装置
CN113273214B (zh) * 2018-12-19 2023-08-29 Lg电子株式会社 基于二次变换的图像编码方法及其装置
US11968397B2 (en) 2018-12-19 2024-04-23 Lg Electronics Inc. Video coding method on basis of secondary transform, and device for same

Also Published As

Publication number Publication date
EP2493197A1 (en) 2012-08-29
US20120201303A1 (en) 2012-08-09
KR20120060914A (ko) 2012-06-12
BR112012011325A2 (pt) 2016-04-19
AU2010310286B2 (en) 2014-05-01
AU2010310286A1 (en) 2012-06-14
CN102045560B (zh) 2013-08-07
US9723313B2 (en) 2017-08-01
EP2493197A4 (en) 2014-07-16
CN102045560A (zh) 2011-05-04
KR101481642B1 (ko) 2015-01-22
BR112012011325B1 (pt) 2019-04-30

Similar Documents

Publication Publication Date Title
WO2011047579A1 (zh) 一种视频编解码方法及设备
KR101626006B1 (ko) 콜로케이티드 영상을 이용한 인터 예측을 수반하는 비디오 부호화 방법 및 그 장치, 비디오 복호화 방법 및 그 장치
JP5767387B2 (ja) ビデオ符号化方法及び装置、ビデオ復号化方法及び装置
JP5824148B2 (ja) 単一化された参照可能性確認過程を介してイントラ予測を伴うビデオ符号化方法及びその装置、ビデオ復号化方法及びその装置
KR20110083367A (ko) 계층적 데이터 단위의 패턴 정보를 이용하는 비디오 부호화 방법과 그 장치, 및 비디오 복호화 방법과 그 장치
KR20110083366A (ko) 스킵 및 분할 순서를 고려한 비디오 부호화 방법과 그 장치, 및 비디오 복호화 방법과 그 장치
KR20110010324A (ko) 영상의 부호화 방법 및 장치, 영상 복호화 방법 및 장치
KR20110017720A (ko) 적응적인 루프 필터링을 이용한 비디오의 부호화 방법 및 장치, 비디오 복호화 방법 및 장치
WO2011153888A1 (zh) 编码方法、解码方法、编码装置、解码装置及编解码系统

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: 10824420

Country of ref document: EP

Kind code of ref document: A1

WWE Wipo information: entry into national phase

Ref document number: 3557/CHENP/2012

Country of ref document: IN

NENP Non-entry into the national phase

Ref country code: DE

WWE Wipo information: entry into national phase

Ref document number: 2010824420

Country of ref document: EP

ENP Entry into the national phase

Ref document number: 20127011936

Country of ref document: KR

Kind code of ref document: A

WWE Wipo information: entry into national phase

Ref document number: 2010310286

Country of ref document: AU

ENP Entry into the national phase

Ref document number: 2010310286

Country of ref document: AU

Date of ref document: 20100830

Kind code of ref document: A

REG Reference to national code

Ref country code: BR

Ref legal event code: B01A

Ref document number: 112012011325

Country of ref document: BR

ENP Entry into the national phase

Ref document number: 112012011325

Country of ref document: BR

Kind code of ref document: A2

Effective date: 20120420