WO2014173092A1 - 一种图像预测编码方法及图像编码器 - Google Patents
一种图像预测编码方法及图像编码器 Download PDFInfo
- Publication number
- WO2014173092A1 WO2014173092A1 PCT/CN2013/085426 CN2013085426W WO2014173092A1 WO 2014173092 A1 WO2014173092 A1 WO 2014173092A1 CN 2013085426 W CN2013085426 W CN 2013085426W WO 2014173092 A1 WO2014173092 A1 WO 2014173092A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- tile
- current frame
- frame
- lcu
- reference frame
- Prior art date
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/50—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/50—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding
- H04N19/503—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding involving temporal prediction
- H04N19/51—Motion estimation or motion compensation
- H04N19/513—Processing of motion vectors
- H04N19/517—Processing of motion vectors by encoding
- H04N19/52—Processing of motion vectors by encoding by predictive encoding
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/102—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or selection affected or controlled by the adaptive coding
- H04N19/103—Selection of coding mode or of prediction mode
- H04N19/105—Selection of the reference unit for prediction within a chosen coding or prediction mode, e.g. adaptive choice of position and number of pixels used for prediction
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/134—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or criterion affecting or controlling the adaptive coding
- H04N19/167—Position within a video image, e.g. region of interest [ROI]
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/169—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding
- H04N19/17—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding the unit being an image region, e.g. an object
- H04N19/172—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding the unit being an image region, e.g. an object the region being a picture, frame or field
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/169—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding
- H04N19/17—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding the unit being an image region, e.g. an object
- H04N19/174—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding the unit being an image region, e.g. an object the region being a slice, e.g. a line of blocks or a group of blocks
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/169—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding
- H04N19/182—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding the unit being a pixel
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/42—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals characterised by implementation details or hardware specially adapted for video compression or decompression, e.g. dedicated software implementation
- H04N19/436—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals characterised by implementation details or hardware specially adapted for video compression or decompression, e.g. dedicated software implementation using parallelised computational arrangements
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/50—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding
- H04N19/503—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding involving temporal prediction
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/65—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using error resilience
Definitions
- the present invention relates to the field of image processing technologies, and in particular, to an image predictive encoding method and an image encoder. Background technique
- HEVC High Efficiency Video Codec/H.265
- HEVC High Efficiency Video Codec/H.265
- It is a hybrid video coding standard based on combination of transform and prediction. It is a chip that is very suitable for parallelization and introduces many parallels.
- the Tile technology divides each frame of image into several rectangular areas, and each rectangular area is divided into a Tile, as shown in Figure 1.
- Each Tile in FIG. 1 includes an integer number of LCUs (Large Coding Units), and each LCU in the Tile is separately encoded in LCU units for image encoding, since each LCU in each Tile is independently performed.
- the encoding is such that each tile is a set of independent coding units, which breaks the dependencies in the space, so that each tile divided in each frame is independent of each other, and several Tiles can be simultaneously encoded to implement parallel processing.
- each PU is an image block including several pixels, when performing PU prediction.
- the reference pixel needs to be determined, and the PU is predicted by using the MV (Moving Vector) of the reference pixel, and the existing predictive coding method includes an intra prediction coding method and an inter prediction coding method, and performing intra prediction coding.
- the selected reference pixel is only related to the Tile of the LCU currently being predicted to be encoded, and has nothing to do with other Tiles.
- the current frame may use a pixel in any Tile in the reference frame as a reference pixel.
- the image includes four Tiles, namely TileO, Tilel, and Tile2.
- Tile3 when performing a PU prediction of the LCU in Tile2, it is possible to use the pixel points in Tilel and Tile2 as reference pixel points, and then use the MV of the selected reference pixel to perform prediction, therefore, if the reference frame is within the Tilel An error occurs in the image data.
- the error may be spread to Tile2, and the pixels in Tile2 may be used as reference pixels for PU prediction in other Tiles, and so on, which may cause Tilel to occur. The error is spread within the entire frame of the image, thereby reducing the accuracy of the image encoding.
- the PMV Predicted Moving Vector
- the MV of the reference pixel may be predicted according to the determined PMV, and each MV corresponds to one reference pixel, and then the reference pixel may be determined.
- the time domain candidate motion vector in other Tiles may be used as the PMV of the PU to be predicted, for example, when performing Tile2 in FIG.
- the time domain candidate motion vector in the Tile3 may be used as the PMV of the PU included in the LCU to be predicted. Therefore, if the image data in the reference frame within the frame is erroneous, the interframe predictive coding is performed. It is possible that the 4 bar error spreads to Tile2, and the time domain candidate motion vector in Tile2 is likely to be the PMV when PU prediction is performed in other Tiles, and so on, which may cause the error of Tile3 to spread in the entire frame image. , thereby reducing the accuracy of image encoding. Summary of the invention
- Embodiments of the present invention provide an image predictive encoding method and an image encoder to prevent errors in image data generated in a Tile from spreading in an entire frame image, thereby improving image encoding accuracy.
- an image predictive coding method including:
- Predictive coding is performed on the LCU according to a prediction result obtained by predicting the PU.
- selecting a reference pixel in the determined tile of the reference frame includes:
- the position in the Tile is the same as or adjacent to the location of the LCU to be predictively encoded in the current frame.
- an image predictive coding method including:
- the method further includes:
- an image encoder including a receiving unit for receiving an image to be encoded, further comprising: a determining unit, a predicting unit, and an encoding unit, where
- the determining unit is configured to obtain the received image information to be encoded from the receiving unit, determine a current frame coding block Tile where the maximum coding unit LCU to be predictively encoded is located in the current frame, and a reference frame in the current frame. Determining a Tile of a reference frame corresponding to a Tile of the current frame, a Tile of a reference frame corresponding to a Tile of the current frame, a position in the reference frame, and a Tile of the current frame in the current frame The positions in the same are the same, and the determined position information of the Tile of the reference frame is output to the prediction unit;
- the prediction unit is configured to acquire location information of a tile of the reference frame determined by the determining unit, and select a reference pixel in the determined Tile of the reference frame, and select the reference pixel point according to the reference pixel point in the LCU
- the included prediction unit PU performs prediction ', outputs the prediction result to the coding unit; predicts the result, and performs predictive coding on the LCU according to the prediction result obtained by predicting the PU.
- the predicting unit is specifically configured to: determine, according to a location of the LCU of the current frame, according to the position of the LCU to be predictively encoded.
- the reference pixel is selected within the tile, wherein the location of the selected LCU to which the reference pixel belongs belongs in the tile of the reference frame is the same as or adjacent to the position of the LCU of the to-be-predicted code in the current frame.
- an image encoder including a receiving unit for receiving an image to be encoded, further comprising: a determining unit, a predicting unit, and an encoding unit, where
- the determining unit is configured to obtain the received image information to be encoded from the receiving unit, determine a current frame coding block Tile where the maximum coding unit LCU to be predictively encoded is located in the current frame, and a reference frame in the current frame. Determining a Tile of a reference frame corresponding to a Tile of the current frame, a Tile of a reference frame corresponding to a Tile of the current frame, a position in the reference frame, and a Tile of the current frame in the current frame The positions in the same are the same, and the determined position information of the Tile of the reference frame is output to the prediction unit;
- the prediction unit is configured to acquire location information of a tile of the reference frame determined by the determining unit, and select a time domain candidate motion vector within the determined Tile of the reference frame as a prediction unit PU included in the LCU Predicting a motion vector PMV, obtaining a motion vector MV of the PU according to the PMV prediction, and predicting the PU according to the predicted MV, and outputting the prediction result to the coding unit; Predicting the result, and performing prediction encoding on the LCU according to the prediction result obtained by predicting the PU.
- the prediction unit is specifically configured to: select a time domain candidate motion vector in a reference frame of a current frame;
- a reference corresponding to the Tile of the current frame in which the LCU to be predicted is located is determined within the reference frame of the current frame.
- the position of the reference frame in the reference frame is the same as the position of the current frame in the current frame, and the reference pixel is selected within the determined reference frame, so that the current predictive coded LCU performs In the PU prediction, the pixel in the reference frame of the reference frame corresponding to the current frame in the reference frame is used as the reference pixel, and the pixel in the other tiles in the reference frame is not used as the reference pixel, so that the reference frame is used as the reference frame. If the other Tiles have an error, the Tile of the current frame where the predictive coded LCU is located will not be incorrect. Similarly, when the current frame of the LCU in which the predictive code is located has an error, it will only limit its error to the Tile of the current frame. Inside, it does not affect other tiles, so it can prevent the error of a certain tile from spreading in the entire frame image. High accuracy of image coding.
- a reference corresponding to the Tile of the current frame in which the LCU to be predicted is located is determined within the reference frame of the current frame.
- the PMV so that when determining the PMV of the PU, the motion vector in the reference frame of the reference frame corresponding to the location of the current frame of the current frame in which the LCU is to be predicted in the reference frame is used, and the motion vectors in other Tiles in the reference frame are not used, so that When an error occurs in a tile, the error is always limited to the area corresponding to the Tile, and the encoding of other Tiles is not affected, so that the error does not spread within the entire frame image, thereby improving the accuracy of image encoding.
- FIG. 1 is a schematic diagram of dividing an image into a tile in the prior art
- FIG. 2 is a schematic diagram of interframe prediction coding in the prior art
- FIG. 3 is a flowchart of an image predictive encoding method according to Embodiment 1 of the present invention.
- FIG. 4A is a schematic diagram of dividing an image Tile according to Embodiment 2 of the present invention.
- 4B is a schematic diagram of interframe prediction coding according to Embodiment 2 of the present invention.
- FIG. 5 is a flowchart of an image predictive coding method according to Embodiment 3 of the present invention.
- FIG. 6 is a schematic diagram of selecting a time domain candidate motion vector in a Tile of a reference frame according to Embodiment 4 of the present invention. Flowchart for implementing PMV for PU;
- FIG. 7 is a schematic diagram of maintaining independence between tiles according to an embodiment of the present invention.
- FIG. 8 is a schematic structural diagram of an image encoder according to an embodiment of the present invention. detailed description
- the image predictive coding method provided by the present invention first determines the Tile of the current frame in which the LCU to be predictively encoded is located in the current frame of the current frame, and then in the current frame. a tile of a reference frame whose position is determined to be the same as the position of the current frame in the current frame, and when the PU included in the LCU is predicted, the reference pixel is selected within the Tile of the determined reference frame; In the case of PMV, the time domain candidate motion vector is selected as the PMV of the PU within the determined reference frame, so that in the LCU encoding process, the pixel within the Tile of the reference frame corresponding to the Tile of the current frame in which the LCU is to be predicted is referenced is referenced.
- a reference pixel refers to the motion vector in the Tile of the reference frame corresponding to the Tile of the current frame in which the LCU is to be predicted to be used as the PMV of the PU, so that when an error occurs in a certain tile in the reference frame, the error is always limited to In the area where the error should occur, it is possible to prevent the error of a certain tile from spreading in the entire frame image, thereby effectively improving the area.
- the accuracy of image coding refer to the motion vector in the Tile of the reference frame corresponding to the Tile of the current frame in which the LCU is to be predicted to be used as the PMV of the PU
- Embodiment 1 is a diagrammatic representation of Embodiment 1:
- the inter-predictive coding of the LCU in the Tile needs to be performed.
- the most important one is to select the reference pixel of the PU in the LCU to be predicted in the reference frame of the current frame. Point, and then perform PU prediction according to the selected reference pixel, and perform inter prediction coding on the LCU in the Tile according to the prediction result obtained by predicting the PU.
- the embodiment of the present invention provides an image prediction coding method, as shown in the figure. 3, including:
- S101 Determine a Tile of a current frame where the LCU to be predictively encoded in the current frame is located.
- the LCU is used as the basic coding unit defined in HEVC/H.265.
- each LCU in each tile is separately encoded in the order of Tile, and then the code is obtained.
- the LCUs form a Tile stream for transmission.
- the tile of the current frame where the LCU to be predictively encoded in the current frame is located is determined in advance, so as to be able to determine, in the reference frame of the current frame, the current frame corresponding to the LCU to be predictively encoded.
- the Tile of the reference frame is determined in advance, so as to be able to determine, in the reference frame of the current frame, the current frame corresponding to the LCU to be predictively encoded.
- S102 Determine a Tile of the reference frame corresponding to the Tile of the current frame determined in S101 within the reference frame of the current frame.
- the tile of each frame is fixed, and the position of each tile is also fixed after the division, which is selected in the reference frame in the embodiment of the present invention.
- the position of the reference frame's Tile in the reference frame is the same as the position of the current frame's Tile in the current frame.
- Tile technology divides each frame image into rows and columns by using vertical and horizontal boundaries. These divided rows and columns respectively form a plurality of rectangular regions, and each rectangular region is a Tile, and is transmitted in the Tile stream.
- the position of the tile is fixed in the process. Therefore, in the embodiment of the present invention, when the tile of the reference frame corresponding to the current frame is determined in the reference frame of the current frame, the current frame may be based on the current frame. The position in the frame, the position selected in the reference frame of the current frame is the same as the position of the current frame in the current frame.
- the Tile with the same position as the Tile of the current frame in the current frame may be selected as the final required in the reference frame of the current frame according to the number of the Tile.
- the Tile of the reference frame may be selected as the final required in the reference frame of the current frame according to the number of the Tile.
- S103 Select a reference pixel in the Tile of the reference frame determined in S102.
- the reference frame of the reference frame determined in S103 is selected, and is not selected in other Tiles in the reference frame to ensure the current frame is determined and determined.
- the Tile-related reference frame is independent of other Tiles in the reference frame, and maintains the independence between the current frame's Tile and other Tiles in the reference frame to avoid the error of a certain Tile in the reference frame from spreading throughout the entire frame. Thereby effectively improving the accuracy of image coding.
- S104 Perform prediction on the PU included in the LCU by using the reference pixel selected in S103, and perform predictive coding on the LCU to be predictively coded according to the obtained prediction result.
- each LCU is divided into a plurality of PUs in the process of performing coding.
- reference pixel points are respectively selected in the Tile of the reference frame, and the selected reference pixel points are used.
- the PU performs prediction, and obtains a predicted value of the PU, and performs a difference operation between the obtained predicted value and the original value of the PU to obtain a residual value of the PU, and predictively encodes the residual value of each PU to complete the pair. Predictive coding of the LCU.
- the position of the current frame in the current frame is determined to be the same as the position of the current frame in the current frame of the LCU to be predictively encoded.
- the Tile is used as the Tile of the reference frame, and selects the reference pixel of the PU included in the LCU to be predicted in the determined reference frame, so that when the PU prediction is performed, the reference frame of the current frame corresponds to the Tile of the current frame.
- the pixel in the Tile of the reference frame is used as the reference pixel, and the pixel in other Tiles in the reference frame is not used as the reference pixel.
- the LCU of the predictive encoding is not located.
- the tile has an error.
- the current frame of the LCU in which the encoded LCU is located is faulty, it will only limit its error to the current frame's Tile. It will not affect other Tiles, and thus prevent a Tile error from occurring. Diffusion within the frame image, thus improving the accuracy of image coding.
- Embodiment 2 is a diagrammatic representation of Embodiment 1:
- the second embodiment of the present invention will be described in detail with reference to the actual application, and the image predictive coding method in the first embodiment is described in detail.
- the second embodiment of the present invention as a preferred embodiment of the first embodiment, is in the Tile of the determined reference frame in S103.
- the process of selecting the reference pixel of the PU is described in detail, other steps and examples The same, will not be described here.
- each image frame is divided into four Tiles, namely TileO, Tilel, Tile2, and Tile3, each of which contains an integer number of LCUs, and each LCU is divided into several PUs, as shown in FIG. 4A.
- each PU in each LCU needs to be predicted separately, and then the predicted PU is used to perform predictive coding on each LCU in the Tile.
- the prediction coding of the LCU 201 is described as an example.
- the current frame of the current frame is the Tile2 of the current frame
- the Tile of the current frame is the Tile of the reference frame in the reference frame of the current frame.
- the pixel is randomly selected as the reference pixel of the PU within the Tile2 of the reference frame.
- the position of the divided Tile in each frame is fixed, the position of the LCU in each tile is also fixed.
- the reference pixel when the reference pixel is selected, in order to improve the accuracy of the prediction coding, Predicting the position of the encoded LCU in the current frame's Tile, selecting the reference pixel within the Tile of the determined reference frame, so that the position of the LCU to which the selected reference pixel belongs belongs in the Tile of the reference frame, and the code to be predicted
- the position of the LCU in the current frame is the same or adjacent. For example, in FIG. 2, when the LCU 201 is encoded, the LCU 201 may be selected and coded in the Tile2 of the reference frame according to the position of the LCU 201 of the current frame.
- the image points in the same or adjacent LCUs are used as reference pixel points.
- the pixels in the LCU 202 and the LCU 203 may be selected as reference pixel points.
- the position of the LCU 202 in the Tile2 of the reference frame is the same as the position of the LCU 201 in the current frame, and the position of the LCU 203 in the Tile2 of the reference frame and the L CU201 is adjacent to the position of Tile2 of the current frame.
- the reference pixel in the reference frame of the current frame is selected from the reference frame of the reference frame of the current frame at the position of the current frame, so that the current prediction encoded LCU performs PU prediction, and the reference frame corresponds to the current frame.
- the pixel in the Tile of the reference frame of the Tile is used as the reference pixel, and the pixel in other Tiles in the reference frame is not used as the reference pixel, which can prevent the error of a certain Tile from spreading in the entire frame to improve the image.
- the accuracy of the coding is selected.
- the reference pixel is selected according to the actual situation according to the position of the current frame of the LCU in which the LCU is to be predicted to be encoded, so that the LCU to be predicted can be further accurately predicted. Further improve the accuracy of image coding.
- Embodiment 3 is a diagrammatic representation of Embodiment 3
- the third embodiment of the present invention provides an image predictive coding method. As shown in FIG. 5, the method includes:
- S301 Determine, in the received current frame, a Tile of a current frame in which the LCU to be predictively encoded in the current frame is located.
- determining the current frame of the current frame in which the LCU to be predicted and encoded in the current frame is located may be determined by using the method in Embodiment 1, and is no longer ⁇ Said.
- S302 Determine a Tile of the reference frame corresponding to the Tile of the current frame determined in S301 within the reference frame of the current frame.
- the tile of the reference frame corresponding to the tile of the current frame determined in the reference frame of the current frame is in the same position as the tile of the current frame in the current frame.
- S303 Select a time domain candidate motion vector in the Tile of the reference frame determined in S302 as the PMV of the PU.
- the selection is performed in the Tile of the reference frame determined in S303, and is not selected in other Tiles in the reference frame to ensure the Tile of the current frame and when
- S304 And predicting the PU according to the predicted MV, and performing predictive coding on the LCU to be predictively coded according to the prediction result obtained by predicting the PU.
- the MV of the PU is predicted by using the difference between the determined PMV and the current vector, and the PU is predicted according to the predicted MV, and the PU is predicted according to the prediction.
- the prediction result is predictively encoded by the LCU that is predictively coded.
- determining, in the reference frame of the current frame, a tile whose position is the same as the position of the current frame in the current frame in which the LCU of the to-be-predicted encoding is located in the current frame is used as the tile of the reference frame, and
- the time domain candidate motion vector is selected as the PMV of the PU in the determined reference frame, so that when determining the PMV of the PU, the time domain candidate motion in the reference frame of the reference frame corresponding to the current frame of the current frame in which the LCU is to be predicted in the reference frame is used.
- Embodiment 4 is a diagrammatic representation of Embodiment 4:
- the fourth embodiment of the present invention will be described in detail with reference to the actual application.
- the fourth embodiment of the present invention is a preferred embodiment of the third embodiment, and is in the tile of the determined reference frame in S303.
- the process of selecting the time domain candidate motion vector as the PMV of the PU is described in detail.
- the other steps are the same as those in the third embodiment, and are not described here.
- the motion vector spatial domain candidate motion vector and the time domain candidate motion vector used in the image encoding process when the candidate motion vector is used to calculate the PMV, the spatial domain candidate motion vector may be selected in the current frame according to the spatial candidate motion vector of the current frame, as the PU PMV, the time domain candidate motion vector can also be selected in the reference frame as the PMV of the PU.
- the spatial candidate motion vector represents the spatial location of the LCU in the current frame.
- the current selection may be selected.
- the spatial candidate motion vectors of other LCUs in the Tile's Tile are used as PMVs instead of the spatial candidate motion vectors in other Tiles to maintain the independence between the Tiles.
- the PU is determined.
- the PMV without selecting the time domain candidate motion vector to determine the PMV of the PU, maintains the independence of the Tile between frames.
- the following parameters are specified in the PPS:
- the time domain candidate motion vector is selected in the tile of the determined reference frame in the embodiment of the present invention.
- the implementation process of the PMV as the PU includes:
- S401 Select a time domain candidate motion vector in a reference frame of the current frame.
- a time domain candidate motion vector when the time domain candidate motion vector is selected, a time domain candidate motion vector may be randomly selected in the reference frame of the current frame, and after the time domain candidate motion vector is selected,
- S402 Determine whether the time domain candidate motion vector selected in S401 is within the Tile of the reference frame corresponding to the Tile of the current frame, and if the selected time domain candidate motion vector is within the Tile of the determined reference frame, then go to S403; If the time domain candidate motion vector is not within the Tile of the determined reference frame, then step S404 is performed.
- the time domain candidate motion vector of each PU in different Tiles is different, and may be according to a time domain in a reference frame corresponding to a Tile of a current frame in which the PU is to be predicted in the current frame.
- the candidate motion vector determines whether the time domain candidate motion vector selected in S401 is within the Tile of the reference frame corresponding to the Tile of the current frame.
- S403 The time domain candidate motion vector selected by S401 is used as the PMV of the PU included in the coded LCU to be predicted.
- the time domain candidate selected by S401 may be selected.
- Motion vector as the PMV of the PU.
- S404 Continue to select the time domain candidate motion vector in the reference frame of the current frame, return to S402, and repeat the above steps until the selected time domain candidate motion vector corresponds to the Tile of the current frame in which the LCU is to be predicted in the current frame.
- the reference frame is inside the Tile.
- the time domain candidate motion vector is selected as the PMV of the PU in the Tile of the reference frame corresponding to the Tile of the current frame in which the current frame is to be predicted in the current frame, so that the PM is used for PU.
- the time domain candidate motion vector in the reference frame of the reference frame with the same position as the current frame is selected, and the time domain candidate motion vector of other Tiles in the reference frame is not selected, so that the inter-frame is The independence of each tile between frames is preserved when predicting encoding.
- Embodiment 5 of the present invention provides an implementation method for maintaining independence between tiles in an intra prediction encoding process and an interframe prediction encoding process, as shown in FIG.
- the intra parameters include the derivation of the PMV, the intra prediction mode derivation, QP (Quantization) Parameter, quantization parameter) derivation, etc.
- the PMV is derived using the spatial candidate motion vector
- the spatial candidate motion vector of the adjacent position across the Tile is not referred to, thereby maintaining the independence between the Tiles; also performing intra prediction
- the luminance prediction mode across the Tile is not referenced to maintain the independence between the Tiles.
- the QP quantization group QP across the Tile cannot be referenced, and the QPs of other Tiles cannot be assigned to the QP to maintain the Tile. The independence between.
- the intra prediction value when the intra prediction value is derived, the sample points across the Tile cannot be referenced, and the independence between the Tiles in the frame is maintained.
- each Tile uses a separate CABAC probability model to maintain independence between tiles within the frame.
- the time domain candidate motion vector in the Tile of the reference frame with the same position as the current frame to be predicted and encoded in the reference frame of the current frame is used as the PMV of the PU, and each frame is maintained. Independence between tiles.
- each tile in the predictively encoded image can maintain independence not only in the frame, but also maintain independence between frames, when an error occurs in a tile, whether in coding.
- Intra-prediction encoding is performed in the current frame, or the current frame obtained by encoding is used as a reference frame for inter-frame prediction encoding, and the error generated by the Tile is limited to the area corresponding to the Tile, and is not spread in the entire frame image, so The accuracy of image coding is better improved.
- an embodiment of the present invention provides an image encoder, where the encoder includes a receiving unit 1 for receiving an image to be encoded, and further includes: determining unit 2 , the prediction unit 3 and the coding unit 4, as shown in FIG. 8, wherein
- a determining unit 2 configured to acquire the received image information to be encoded from the receiving unit 1, determine a current frame coding block Tile where the maximum coding unit LCU to be predictively encoded is located in the current frame; and determine in a reference frame of the current frame
- the Tile of the reference frame corresponding to the Tile of the current frame, the Tile of the reference frame corresponding to the Tile of the current frame, the position in the reference frame is the same as the position of the current frame in the current frame, and the determined reference is The position information of the Tile's Tile is output to the prediction unit 3.
- a prediction unit 3 configured to acquire location information of a Tile of a reference frame determined by the determining unit 2, select a reference pixel in a Tile of the determined reference frame, and select, according to the reference pixel point, the LCU
- the prediction unit PU performs prediction and outputs the prediction result to the encoding unit 4.
- the coding unit 4 is configured to obtain a prediction result obtained by predicting the PU output by the prediction unit 3, and perform prediction coding on the LCU according to the prediction result obtained by predicting the PU.
- forecasting unit 3 is specifically used for:
- the reference pixel in the reference frame of the current frame is selected from the reference frame of the reference frame of the current frame at the position of the current frame, so that the current prediction encoded LCU performs PU prediction, and the reference frame corresponds to the current frame.
- the pixel in the Tile of the reference frame of the Tile is used as the reference pixel, and the pixel in other Tiles in the reference frame is not used as the reference pixel, which can prevent the error of a certain Tile from spreading in the entire frame to improve the image.
- the accuracy of the coding when the reference pixel is selected, the reference pixel is selected according to the actual situation according to the position of the current frame of the LCU in which the LCU is to be predicted to be encoded, so that the LCU to be predicted can be further accurately predicted. Further improve the accuracy of image coding.
- the embodiment of the present invention provides an image encoder, which includes a receiving unit 1 for receiving an image to be encoded, and further includes: a determining unit 2, based on the image predictive encoding and decoding method of the third embodiment and the fourth embodiment. , the prediction unit 3 and the coding unit 4, wherein
- a determining unit 2 configured to acquire the received image information to be encoded from the receiving unit 1, determine a current frame coding block Tile where the maximum coding unit LCU to be predictively encoded is located in the current frame; and determine in a reference frame of the current frame a Tile of a reference frame corresponding to a Tile of the current frame, a Tile of a reference frame corresponding to a Tile of the current frame, a position in the reference frame is the same as a position of a current frame of 1 in the current frame, and The position information of the determined Tile of the 1 reference frame is output to the 1 prediction unit 3.
- a prediction unit 3 configured to acquire location information of a Tile of a reference frame determined by the determining unit 2, and select a time domain candidate motion vector within a Tile of the determined reference frame as a predicted motion vector PMV of the prediction unit PU included in the LCU,
- the motion vector MV of the PU is obtained according to the PMV prediction, and the PU is predicted based on the predicted MV, and the prediction result is output to the coding unit 4.
- the coding unit 4 is configured to obtain a prediction result obtained by predicting the PU output by the prediction unit 3, and perform prediction coding on the LCU according to the prediction result obtained by predicting the PU.
- forecasting unit 3 is specifically used for:
- the selected time domain candidate motion vector is used as the PMV of the PU included in the LCU; if not, the process of continuing to select the time domain candidate motion vector in the reference frame of the current frame is returned.
- determining the PMV of the PU determining, in the reference frame of the current frame, a tile whose position is the same as the position of the current frame in the current frame in which the LCU of the to-be-predicted encoding is located in the current frame is used as the tile of the reference frame, and Selecting a time domain candidate motion vector within the Tile of the determined reference frame as the PMV of the PU,
- the time domain candidate motion vector in the reference frame of the reference frame of the current frame in which the LCU is to be predicted in the reference frame is used, and the time domain candidate motion vector of other Tiles in the reference frame is not used, so that when When an error occurs in a tile, the error is always limited to the area corresponding to
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Computing Systems (AREA)
- Theoretical Computer Science (AREA)
- Compression Or Coding Systems Of Tv Signals (AREA)
- Compression, Expansion, Code Conversion, And Decoders (AREA)
Abstract
提供了一种图像预测编码方法及编码器。在进行当前帧的Tile内LCU编码过程中,首先确定待预测编码的LCU或PU所在当前帧的Tile,然后在当前帧的参考帧内确定位置与当前帧的Tile在当前帧中的位置相同的参考帧的Tile,在对LCU的PU进行预测时,在确定的参考帧的Tile内选择参考像素点,在确定LCU包含的PU的PMV时,在确定的参考帧的Tile内选择PU的PMV,能够防止某个Tile发生的错误在整帧图像内扩散,从而提高了图像编码的精确性。
Description
一种图像预测编码方法及图像编码器 本申请要求于 201 3年 4月 16日提交中国专利局、 申请号为 201 310149199. X、 发明名称为"一种图像预测编码方法及图像编码器"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
本发明涉及图像处理技术领域, 尤其涉及一种图像预测编码方法及图像编 码器。 背景技术
HEVC(High Efficiency Video Codec)/H.265作为当前主流的视频编码标准, 是一种基于变换和预测相结合的混合视频编码标准, 其为适应并行化运算程度 非常高的芯片, 引入了很多并行运算的优化思路, 其中 Tile (编码区块)技术 是一种非常典型的用于实现并行化运算的技术。
Tile技术是将每一帧图像划分为若干个矩形区域, 划分出的每个矩形区域 为一个 Tile, 如图 1所示。 图 1中每一个 Tile包含整数个 LCU ( Largest Coding Unit, 最大编码单元), 进行图像编码时以 LCU为单位, 对 Tile内的每个 LCU 分别进行编码, 由于各个 Tile中的每个 LCU独立进行编码, 使得每个 Tile都 是独立编码单元的集合, 其将空间中的依赖关系打破, 使得每一帧中划分的各 个 Tile之间相互独立 , 可以对几个 Tile同时进行编码, 实现并行处理。
现有技术中,对 LCU进行编码时, 需要对每个 LCU内包含的至少一个 PU ( Predicting Unit, 预测单元)进行预测, 每个 PU为包含若干个像素点的图像 块, 在进行 PU预测时, 需要确定参考像素点, 利用参考像素点的 MV ( Moving Vector, 运动矢量)对 PU进行预测, 而现有的预测编码方法包括帧内预测编码 方法和帧间预测编码方法, 进行帧内预测编码时, 由于帧内各个 Tile之间的独 立性, 进行 PU预测时, 选择的参考像素点只与当前待预测编码的 LCU所在的 Tile有关, 而与其他 Tile无关。
但在进行帧间预测编码过程中, 一方面, 当前帧可能将参考帧内任一 Tile 内的像素点作为参考像素点,如图 2所示,图像包括 4个 Tile分别为 TileO、 Tilel、 Tile2和 Tile3 , 当进行 Tile2中 LCU的某个 PU预测时, 可能使用 Tilel和 Tile2 中的像素点作为参考像素点, 然后利用选择的参考像素点的 MV进行预测, 因 此, 如果参考帧内 Tilel内的图像数据发生错误, 经过帧间预测编码, 可能把错 误扩散至 Tile2 , 而 Tile2中的像素点又有可能作为其他 Tile内进行 PU预测时 的参考像素点,以此类推,可能就会使 Tilel发生的错误在整帧图像内进行扩散, 从而降低了图像编码的精确性。
另一方面, 进行 PU预测时, 在确定参考像素点时, 最主要的是要确定 PU 的 PMV ( Predicted Moving Vector , 预测运动矢量), 当确定了 PU的 PMV后,
可根据确定的 PMV对参考像素点的 MV进行预测, 而每个 MV对应一个参考 像素点, 进而可以确定参考像素点。 但是在进行帧间预测时, 在参考帧内选择 时域候选运动矢量作为 PU的 PMV时,可能使用其它 Tile内的时域候选运动矢 量作为待预测 PU的 PMV,例如图 2中当进行 Tile2中 LCU的某个 PU预测时, 可能使用 Tile3内的时域候选运动矢量作为待预测编码 LCU内包含的 PU的 PMV, 因此, 如果参考帧内 Tile3内的图像数据发生错误, 经过帧间预测编码, 可能 4巴错误扩散至 Tile2 ,而 Tile2中的时域候选运动矢量又有可能作为其他 Tile 内进行 PU预测时的 PMV, 以此类推, 可能就会使 Tile3发生的错误在整帧图 像内进行扩散, 从而降低了图像编码的精确性。 发明内容
本发明实施例提供一种图像预测编码方法及图像编码器, 以防止某个 Tile 内图像数据发生的错误在整帧图像内扩散, 提高图像编码的精确性。
第一方面, 提供一种图像预测编码方法, 包括:
在接收到的当前帧内,确定待预测编码的最大编码单元 LCU所在的当前帧 编码区块 Tile;
在当前帧的参考帧内确定与所述当前帧的 Tile相对应的参考帧的 Tile, 其 中与所述当前帧的 Tile相对应的参考帧的 Tile, 在参考帧中的位置与所述当前 帧的 Tile在所述当前帧中的位置相同;
在确定的所述参考帧的 Tile内选择参考像素点, 并根据所述参考像素点对 所述 LCU内包含的预测单元 PU进行预测;
根据对所述 PU进行预测得到的预测结果, 对所述 LCU进行预测编码。 结合第一方面, 在第一种可能的实现方式中, 在确定的所述参考帧的 Tile 内选择参考像素点, 包括:
根据所述待预测编码的 LCU在当前帧的 Tile中所处的位置,在确定的所述 参考帧的 Tile内选取参考像素点,其中,选取的所述参考像素点所属的 LCU在 参 帧的 Tile中的位置, 与所述待预测编码的 LCU在当前帧的 Tile中的位置 相同或相邻
第二方面, 提供一种图像预测编码方法, 包括:
在接收到的当前帧内,确定待预测编码的最大编码单元 LCU所在的当前帧 编码区块 Tile;
在当前帧的参考帧内确定与所述当前帧的 Tile相对应的参考帧的 Tile, 其 中与所述当前帧的 Tile相对应的所述参考帧的 Tile, 在参考帧中的位置与所述 当前帧的 Tile在所述当前帧中的位置相同;
在确定的所述参考帧的 Tile内选择时域候选运动矢量,作为所述 LCU内包 含的预测单元 PU的预测运动矢量 PMV;
根据所述 PMV预测得到所述 PU的运动矢量 MV, 并根据预测得到的所述 MV对所述 PU进行预测;
根据对所述 PU进行预测得到的预测结果, 对所述 LCU进行预测编码。
结合第二方面, 在第一种可能的实现方式中, 在当前帧的参考帧内确定了 所述当前帧的 Tile相对应的参考帧的 Tile之后, 该方法还包括:
在当前帧的参考帧内选择一个时域候选运动矢量;
判断选择的时域候选运动矢量是否在与所述当前帧的 Tile对应的所述参考 帧的 Tile内;
若是, 则利用选择的时域候选运动矢量, 作为所述 LCU内包含的 PU的 PMV;
若否,则返回执行继续在当前帧的参考帧内选择时域候选运动矢量的处理。 第三方面, 提供一种图像编码器, 包括用于接收待编码图像的接收单元, 还包括: 确定单元、 预测单元以及编码单元, 其中,
所述确定单元, 用于从所述接收单元获取接收到的待编码图像信息, 在当 前帧内确定待预测编码的最大编码单元 LCU所在的当前帧编码区块 Tile; 以及 在当前帧的参考帧内确定与所述当前帧的 Tile相对应的参考帧的 Tile, 与所述 当前帧的 Tile相对应的参考帧的 Tile, 在参考帧中的位置与所述当前帧的 Tile 在所述当前帧中的位置相同, 并将确定的所述参考帧的 Tile的位置信息输出至 所述预测单元;
所述预测单元, 用于获取所述确定单元确定的参考帧的 Tile的位置信息, 并在确定的所述参考帧的 Tile内选择参考像素点, 并根据所述参考像素点对所 述 LCU内包含的预测单元 PU进行预测', 将预测结果输出至所述编码单元; 预测结果, 并根据对所述 PU进行预测得到的预测结果, 对所述 LCU进行预测 编码。
结合第三方面, 在第一种可能的实现方式中, 所述预测单元, 具体用于: 根据所述待预测编码的 LCU在当前帧的 Tile中所处的位置,在确定的所述 参考帧的 Tile内选取参考像素点,其中,选取的所述参考像素点所属的 LCU在 参 帧的 Tile中的位置, 与所述待预测编码的 LCU在当前帧的 Tile中的位置 相同或相邻。
第四方面, 提供一种图像编码器, 包括用于接收待编码图像的接收单元, 还包括: 确定单元、 预测单元以及编码单元, 其中,
所述确定单元, 用于从所述接收单元获取接收到的待编码图像信息, 在当 前帧内确定待预测编码的最大编码单元 LCU所在的当前帧编码区块 Tile; 以及 在当前帧的参考帧内确定与所述当前帧的 Tile相对应的参考帧的 Tile, 与所述 当前帧的 Tile相对应的参考帧的 Tile, 在参考帧中的位置与所述当前帧的 Tile 在所述当前帧中的位置相同, 并将确定的所述参考帧的 Tile的位置信息输出至 所述预测单元;
所述预测单元, 用于获取所述确定单元确定的参考帧的 Tile的位置信息, 并在确定的所述参考帧的 Tile内选择时域候选运动矢量,作为所述 LCU内包含 的预测单元 PU的预测运动矢量 PMV, 根据所述 PMV预测得到所述 PU的运 动矢量 MV, 并根据预测得到的所述 MV对所述 PU进行预测, 将预测结果输 出至所述编码单元;
预测结果, 并根据对所述 PU进行预测得到的预测结果, 对所述 LCU进行预测 编码。
结合第四方面, 在第一种可能的实现方式中, 所述预测单元, 具体用于: 在当前帧的参考帧内选择一个时域候选运动矢量;
判断选择的时域候选运动矢量是否在与所述当前帧的 Tile对应的所述参考 帧的 Tile内;
若是, 则利用选择的时域候选运动矢量, 作为所述 LCU内包含的 PU的 PMV;
若否,则返回执行继续在当前帧的参考帧内选择时域候选运动矢量的处理。 根据第一方面提供的图像预测编码方法和第三方面提供的图像编码器, 在 进行 LCU预测编码过程中, 在当前帧的参考帧内确定与待预测编码 LCU所在 当前帧的 Tile相对应的参考帧的 Tile, 其中参考帧的 Tile在参考帧中的位置与 当前帧的 Tile在当前帧中的位置相同, 并在确定的参考帧的 Tile内选择参考像 素点, 从而使得当前预测编码的 LCU进行 PU预测时, 将参考帧内对应当前帧 的 Tile所在位置的参考帧的 Tile内的像素点作为参考像素点, 不会将参考帧内 其他 Tile内的像素点作为参考像素点, 这样当参考帧内其他 Tile发生错误时, 不会使预测编码的 LCU所在的当前帧的 Tile发生错误,同样当预测编码的 LCU 所在的当前帧的 Tile发生错误, 也只会将其错误限制在当前帧的 Tile内, 不会 影响其他 Tile, 因此能够防止某个 Tile发生的错误在整帧图像内扩散, 从而提 高了图像编码的精确性。
根据第二方面提供的图像预测编码方法和第四方面提供的图像编码器, 在 进行 PU预测编码过程中, 在当前帧的参考帧内确定与待预测编码 LCU所在当 前帧的 Tile相对应的参考帧的 Tile, 其中参考帧的 Tile在参考帧中的位置与当 前帧的 Tile在当前帧中的位置相同; 并在参考帧的 Tile内选择时域候选运动矢 量作为待预测编码 LCU内包含的 PU的 PMV, 从而使得确定 PU的 PMV时, 使用参考帧内对应待预测编码 LCU所在当前帧的 Tile所在位置的参考帧的 Tile 内的运动矢量, 不使用参考帧内其他 Tile内的运动矢量, 使得当某个 Tile发生 错误时, 始终将错误限制在对应该 Tile的区域内, 不会影响其他 Tile的编码, 从而使错误不会在整帧图像内扩散, 从而提高了图像编码的精确性。 附图说明
图 1为现有技术中图像划分为 Tile的示意图;
图 2为现有技术中帧间预测编码示意图;
图 3为本发明实施例一提供的图像预测编码方法的流程图;
图 4A所示为本发明实施例二提供的图像 Tile划分示意图;
图 4B为本发明实施例二提供的帧间预测编码示意图;
图 5为本发明实施例三提供的图像预测编码方法的流程图;
图 6为本发明实施例四提供的在参考帧的 Tile内选择时域候选运动矢量作
为 PU的 PMV的实现流程图;
图 7为本发明实施例提供的 Tile之间保持独立性的示意图;
图 8为本发明实施例提供的图像编码器构成示意图。 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。
本发明提供的图像预测编码方法,在进行当前帧某个 Tile内 LCU进行编码 过程中 ,首先在接收到的当前帧内确定待预测编码的 LCU所在的当前帧的 Tile, 然后在当前帧的参考帧内确定位置与当前帧的 Tile在当前帧中的位置相同的参 考帧的 Tile, 在对 LCU内包含的 PU进行预测时, 在确定的参考帧的 Tile内选 择参考像素点; 在确定 PU的 PMV时,在确定的参考帧的 Tile内选择时域候选 运动矢量作为 PU的 PMV,从而使得 LCU编码过程中,参考与待预测编码 LCU 所在当前帧的 Tile对应的参考帧的 Tile内的像素点作为参考像素点, 参考与待 预测编码 LCU所在当前帧的 Tile对应的参考帧的 Tile内的运动矢量作为 PU的 PMV, 进而使得当参考帧内某个 Tile发生错误时, 始终将错误限制在对应该发 生错误的 Tile的区域内, 能够防止某个 Tile发生的错误在整帧图像内扩散, 从 而有效提高了图像编码的精确性。
以下将结合具体的实施例以及附图对本发明提供的图像预测编码方法进行 详细的说明, 当然并不引以为限。
实施例一:
在图像预测编码过程中,需要对 Tile内的 LCU进行帧间预测编码,在进行 LCU帧间预测编码过程中, 最主要的是在当前帧的参考帧内选择待预测编码 LCU内 PU的参考像素点, 然后根据选择的参考像素点进行 PU预测, 并根据 对 PU进行预测后得到的预测结果对 Tile内的 LCU进行帧间预测编码,本发明 实施例提供了一种图像预测编码方法, 如图 3所示, 包括:
S101 : 确定当前帧内待预测编码的 LCU所在的当前帧的 Tile。
具体的, LCU作为 HEVC/H.265中定义的基本编码单元, 在进行图像的 预测编码过程中, 需按照 Tile的顺序对每一 Tile内的每一 LCU分别进行编码, 然后将编码得到的多个 LCU组成 Tile码流进行传输。
本发明实施例中在进行 LCU的编码时, 预先确定当前帧内待预测编码的 LCU所在的当前帧的 Tile, 以便能够在当前帧的参考帧内确定与待预测编码的 LCU所在当前帧相对应的参考帧的 Tile。
S102: 在当前帧的参考帧内确定与 S101中确定的当前帧的 Tile相对应的 参考帧的 Tile。
具体的, 釆用 Tile技术进行图像编码过程中, 每一帧图像划分的 Tile是固 定的, 划分后每一 Tile的位置也是固定的, 本发明实施例中在参考帧内选择的
参考帧的 Tile在参考帧中的位置与当前帧的 Tile在当前帧中的位置相同。
Tile技术是利用垂直和水平的边界将每一帧图像划分为一些行和列, 这些 划分出的行和列分别组成多个矩形区域, 而每一个矩形区域为一个 Tile , 并且 在 Tile码流传输过程中, 这些 Tile的位置是固定不变的, 因此, 本发明实施例 中在当前帧的参考帧内确定与当前帧的 Tile相对应的参考帧的 Tile时, 可根据 当前帧的 Tile在当前帧中的位置,在当前帧的参考帧内选择位置与当前帧的 Tile 在当前帧中的位置相同
考帧内确定与当前帧的 Tile相对应的参考帧的 Tile时, 可根据 Tile的编号在当 前帧的参考帧内选择位置与当前帧的 Tile在当前帧中的位置相同的 Tile作为最 终所需的参考帧的 Tile。
S103: - S102中确定的参考帧的 Tile内选取参考像素点。
具体的, 本发明实施例中在参考帧内选取 PU的参考像素点时, 在 S103中 确定的参考帧的 Tile内进行选取, 不在参考帧内其他 Tile内进行选取, 保证当 前帧的 Tile与确定的参考帧的 Tile相关, 与参考帧内其他 Tile无关, 使当前帧 的 Tile与参考帧内其他 Tile之间保持独立性, 以避免参考帧内某个 Tile发生的 错误在整帧图像内扩散, 从而有效提高了图像编码的精确性。
S104: 利用 S103中选取的参考像素点对 LCU内包含的 PU进行预测, 根 据得到的预测结果, 对待预测编码的 LCU进行预测编码。
具体的, 每个 LCU在进行编码过程中会被划分为若干个 PU, 本发明实施 例中针对 LCU内包括的各个 PU, 在参考帧的 Tile内分别选取参考像素点, 利 用选取的参考像素点对 LCU内包含的各个 PU分别进行预测, 然后根据得到的 预测结果对待预测编码的 LCU进行预测编码, 其中根据对 PU进行预测得到的 预测结果对待预测编码的 LCU进行预测编码, 主要是指通过对 PU进行预测, 可得到的 PU的预测值, 将得到的预测值与 PU的原始值做差运算, 得到 PU的 残差值,通过对每一 PU的残差值进行预测编码,即可完成对 LCU的预测编码。
本发明实施例中 , 在进行选取待预测编码 LCU内包含的 PU的参考像素点 时,在当前帧的参考帧内确定位置与待预测编码的 LCU所在当前帧的 Tile在当 前帧中的位置相同的 Tile作为参考帧的 Tile, 并在确定的参考帧的 Tile内选取 待预测编码 LCU内包含的 PU的参考像素点, 使得进行 PU预测时, 将当前帧 的参考帧内对应当前帧的 Tile的参考帧的 Tile内的像素点作为参考像素点, 不 会将参考帧内其他 Tile内的像素点作为参考像素点, 当参考帧内其他 Tile发生 错误时, 不会使预测编码的 LCU所在当前帧的 Tile发生错误, 同样当预测编码 的 LCU所在当前帧的 Tile发生错误, 也只会将其错误限制在当前帧的 Tile内, 不会影响其他 Tile, 进而能够防止某个 Tile发生的错误在整帧图像内的扩散, 因此提高了图像编码的精确性。
实施例二:
本发明实施例二将结合实际应用对实施例一中的图像预测编码方法进行详 细的说明,本发明实施例二作为实施例一的较优实施例,对 S103中在确定的参 考帧的 Tile内选取 PU的参考像素点的过程进行详细说明, 其他步骤与实施例
一相同, 在此不再赘述。
本发明实施例中将每个图像帧划分为 4个 Tile, 分别为 TileO、 Tilel、 Tile2 和 Tile3 , 每个 Tile中包含整数个 LCU, 而每个 LCU又被划分为若干个 PU , 如图 4A所示, 在进行图像预测编码时, 需对每个 LCU内的各个 PU分别进行 预测, 然后利用预测得到的 PU, 对 Tile内的每一 LCU进行预测编码, 本发明 实施例中以进行 Tile2内 LCU201的预测编码为例进行说明。
如图 4B所示, 本发明实施例待预测编码 LCU201所在的当前帧的 Tile为 当前帧的 Tile2, 当前帧的 Tile2在当前帧的参考帧内对应的参考帧的 Tile为参 考帧的 Tile2, 在选取 LCU201内包含的 PU的参考像素点时, 在参考帧的 Tile2 内任意选取像素点作为 PU的参考像素点。
进一步的, 由于每一帧图像中划分后的 Tile位置固定, 每个 Tile内的 LCU 位置也是固定的, 本发明实施例中在选取参考像素点时, 为提高预测编码的准 确性,可根据待预测编码的 LCU在当前帧的 Tile中所处的位置,在确定的参考 帧的 Tile内选取参考像素点,使选取的参考像素点所属的 LCU在参考帧的 Tile 中的位置,与待预测编码的 LCU在当前帧的 Tile中的位置相同或相邻,例如图 2中, 进行 LCU201编码时, 可根据 LCU201在当前帧的 Tile2所处的位置, 在 参考帧的 Tile2内选取与待预测编码 LCU201所处位置相同或相邻的 LCU内的 像 点作为参考像素点, 如图 4B所示, 选取 LCU201内包含的 PU的参考像素 点时, 可选取 LCU202和 LCU203内的像素点作为参考像素点, 其中, LCU202 在参考帧的 Tile2中的位置与 LCU201在当前帧的 Tile2的位置相同, LCU203 在参 帧的 Tile2中的位置与 LCU201在当前帧的 Tile2的位置相邻。 当前帧的参考帧中位置与当前帧的 Tile在当前帧的位置相同的参考帧的 Tile内 选取 PU的参考像素点, 使得当前预测编码的 LCU进行 PU预测时, 将参考帧 内对应当前帧的 Tile的参考帧的 Tile内的像素点作为参考像素点, 不会将参考 帧内其他 Tile内的像素点作为参考像素点, 能够防止某个 Tile发生的错误在整 帧图像内扩散, 以提高图像编码的精确性。 进一步的, 在选取参考像素点时, 根据待预测编码的 LCU所在当前帧的 Tile的位置在参考帧的 Tile内根据实际 情况进行参考像素点的选取, 能够进一步准确的预测得到待预测编码 LCU, 进 一步提高图像编码的精确性。
实施例三:
在图像预测编码过程中, 进行 PU预测过程中, 还需要在当前帧的参考帧 内选取一些参数, 比如 PMV ( Predicted Moving vector, 预测运动矢量), 根据 选择的 PMV对 PU的 MV进行预测 ,根据预测得到的 MV对 PU进行预测 ,根 据预测结果对 Tile内待预测编码的 LCU进行预测编码,本发明实施例三提供一 种图像预测编码方法, 如图 5所示, 包括:
S301 : 在接收到的当前帧内,确定当前帧内待预测编码的 LCU所在的当前 帧的 Tile。
具体的, 本发明实施例中在接收到的当前帧内, 确定当前帧内待预测编码 的 LCU所在的当前帧的 Tile, 可釆用实施例一中的方法进行确定, 在此不再赘
述。
S302: 在当前帧的参考帧内确定与 S301中确定的当前帧的 Tile相对应的 参考帧的 Tile。
本发明实施例中, 在当前帧的参考帧内确定的与当前帧的 Tile相对应的参 考帧的 Tile在参考帧中的位置与当前帧的 Tile在当前帧中的位置相同。
S303: 在 S302中确定的参考帧的 Tile内选择时域候选运动矢量, 作为 PU 的 PMV。
具体的,本发明实施例中在进行时域候选运动矢量的选取时,在 S303中确 定的参考帧的 Tile内进行选取, 不在参考帧内其他 Tile内进行选取, 保证当前 帧的 Tile与 当
具体的, 本发明实施例中当确定了 PU的 PMV后, 利用确定的 PMV与当 前矢量的差值, 预测得到 PU的 MV, 根据预测得到的 MV对 PU进行预测, 并 根据对 PU进行预测得到的预测结果对待预测编码的 LCU进行预测编码。
本发明实施例中, 在确定 PU的 PMV时, 在当前帧的参考帧内确定位置与 待预测编码的 LCU所在当前帧的 Tile在当前帧中的位置相同的 Tile作为参考 帧的 Tile,并在确定的参考帧的 Tile内选取时域候选运动矢量作为 PU的 PMV, 使得确定 PU的 PMV时,使用参考帧内对应待预测编码 LCU所在当前帧的 Tile 的参考帧的 Tile内的时域候选运动矢量, 不使用参考帧内其他 Tile的时域候选 运动矢量, 使得当某个 Tile发生错误时, 始终将错误限制在对应该 Tile的区域 内, 不会影响其他 Tile的预测编码, 从而使错误不会在整帧图像内扩散, 因此 提高了图像编码的精确性。
实施例四:
本发明实施例四将结合实际应用对实施例三中的图像预测编码方法进行详 细的说明,本发明实施例四作为实施例三的较优实施例,对 S303中在确定的参 考帧的 Tile内选择时域候选运动矢量, 作为 PU的 PMV的过程进行详细说明, 其他步骤与实施例三相同, 在此不再赘述。
图像编码过程中用到的运动矢量 空域候选运动矢量和时域候选运动矢 量, 选取候选运动矢量计算 PMV时, 可根据当前帧的空域候选运动矢量在当 前帧内选取空域候选运动矢量, 作为 PU的 PMV, 也可在参考帧内选取时域候 选运动矢量, 作为 PU的 PMV。
A、 选择当前帧的空域候选运动矢量, 作为 LCU内包含的 PU的 PMV: 空域候选运动矢量表征了当前帧的 Tile内 LCU所在的空间位置,本发明实 施例中在计算 PMV时 ,可选择当前帧的 Tile内其他 LCU的空域候选运动矢量 , 作为 PMV, 而不使用其他 Tile内的空域候选运动矢量,保持各个 Tile之间的独 立性。
具体的, 在帧间预测编码过程中, 若选择空域候选运动矢量确定 PU的
PMV,而不选择时域候选运动矢量确定 PU的 PMV,可保持帧间 Tile的独立性。 例如, 在 HEVC/H.265协议中, 规定 PPS中有如下参数:
enable— temporal— mvp— flag, 在编码时, 将此参数的值设置为 0, 则在整个编码 过程中,选择空 候 i£运动矢量,而不选择时域候选运动矢量,作为 PU的 PMV。
B、 在确定的参考帧的 Tile内选择时域候选运动矢量, 作为 LCU内包含的
PU的 PMV:
如图 6所示为本发明实施例中在确定的参考帧的 Tile内选择时域候选运动 矢量, 作为 PU的 PMV的实现过程, 包括:
S401 : 在当前帧的参考帧内选择一个时域候选运动矢量。
具体的, 本发明实施例中在选择时域候选运动矢量时, 可在当前帧的参考 帧内随机选择一个时域候选运动矢量, 当选择了时域候选运动矢量后, 进行
S402。
S402: 判断 S401中选择的时域候选运动矢量是否在与当前帧的 Tile对应 的参考帧的 Tile内, 若选择的时域候选运动矢量在确定的参考帧的 Tile内, 则 转 S403; 若选择的时域候选运动矢量不在确定的参考帧的 Tile内, 则转步骤 S404。
具体的, 每个 PU在不同的 Tile内的时域候选运动矢量是不同的, 可根据 参考帧内与当前帧内待预测编码 PU所在当前帧的 Tile相对应的参考帧的 Tile 内的时域候选运动矢量,判断 S401中选择的时域候选运动矢量是否在与当前帧 的 Tile对应的参考帧的 Tile内。
S403: 利用 S401选择的时域候选运动矢量, 作为待预测编码 LCU内包含 的 PU的 PMV。
具体的,通过 S402中的判断后,得知选择的时域候选运动矢量在确定的与 待预测编码 LCU所在当前帧的 Tile相对应的参考帧的 Tile内, 则可将 S401选 择的时域候选运动矢量, 作为 PU的 PMV。
S404: 继续在当前帧的参考帧内选择时域候选运动矢量, 返回至 S402, 并 重复执行上述步骤, 直至选择的时域候选运动矢量在当前帧内待预测编码 LCU 所在当前帧的 Tile相对应的参考帧的 Tile内。
本发明实施例中, 在确定 PU的 PMV时, 在当前帧内待预测编码所在当前 帧的 Tile相对应的参考帧的 Tile内选用时域候选运动矢量作为 PU的 PMV,使 得釆用 PMV进行 PU的 MV的帧间预测时, 只选用参考帧内与当前帧的 Tile 位置相同的参考帧的 Tile内的时域候选运动矢量, 不选用参考帧内其他 Tile的 时域候选运动矢量, 使得帧间预测编码时保持帧间各个 Tile的独立性。
实施例五
结合本发明实施例一至实施例四, 本发明实施例五提供一种在帧内预测编 码过程和帧间预测编码过程中保持各个 Tile之间独立性的实施方法, 如图 7所 示:
( 1 )、 使用帧内相关参数进行帧内的参数预测时, 保持帧内各个 Tile之间 的独立性。
具体的,帧内参数包括 PMV的推导、帧内预测模式推导、 QP ( Quantization
Parameter, 量化参数)推导等, 具体的, 在使用空域候选运动矢量推导 PMV 时, 不参考跨 Tile的相邻位置的空域候选运动矢量, 从而保持 Tile之间的独立 性; 同样在进行帧内预测模式推导时, 跨 Tile的亮度预测模式也不参考, 以保 持 Tile之间的独立性; QP推导时, 跨 Tile的 QP量化组 QP不能参考, 不能将 其他 Tile的 QP赋值给 QP , 以保持 Tile之间的独立性。
( 2 )、 在做帧内预测值推导时, 保持帧内各 Tile之间的独立性。
具体的, 在做帧内预测值推导时, 跨 Tile的样本点不能参考, 保持帧内各 个 Tile之间的独立性。
( 3 )、 在故 CAB AC ( Context Adaptive Binary Arithmetic Coder, 自适应二 进制算术编码)编码时, 保持帧内各个 Tile之间的独立性。
具体的, 每个 Tile使用一个独立的 CABAC概率模型, 保持帧内各个 Tile 之间的独立性。
( 4 )、 在做帧内边界滤波时, 保持帧内各个 Tile之间的独立性。
具体的, 通过将语法元素 loop— filter— across— tiles— enabled— flag设置为 0, 控制 Tile边界的像素点不使用相邻 Tile ^像素 进行滤波, 而实现帧内各个 Tile之间的独立性。
( 5 )、 使用帧间像素值进行预测时, 保持帧间各个 Tile之间的独立性。 具体的, LCU预测编码过程中, 对 PU进行预测时, 选择参考帧内对应当 前帧的 Tile所在位置的参考帧的 Tile内的像素点作为参考像素点进行预测, 保 持帧间各个 Tile之间的独立性。
( 6 )、 使用帧间参数进行预测时, 保持帧间各个 Tile之间的独立性。
具体的, 在帧间进行 PU的 PMV确定时, 以当前帧的参考帧内与当前帧待 预测编码 Tile位置相同的参考帧的 Tile内的时域候选运动矢量作为 PU的 PMV, 保持帧间各个 Tile之间的独立性。
通过本发明实施例, 使得预测编码得到的图像中各个 Tile之间不仅在帧内 能够保持独立性, 而且在帧间也能保持独立性, 当某个 Tile发生了错误, 无论 是在编码得到的当前帧内进行帧内预测编码, 还是以编码得到的当前帧为参考 帧进行帧间预测编码, Tile发生的错误都会限制在对应该 Tile的区域内 , 不会 在整帧图像内进行扩散, 因此较好地提高了图像编码的精确性。
实施例六
基于与本发明实施例一和实施例二提供的图像编码方法, 本发明实施例提 供了一种图像编码器, 该编码器包括用于接收待编码图像的接收单元 1 , 还包 括: 确定单元 2、 预测单元 3以及编码单元 4, 如图 8所示, 其中,
确定单元 2, 用于从接收单元 1获取接收到的待编码图像信息, 在当前帧 内确定待预测编码的最大编码单元 LCU所在的当前帧编码区块 Tile; 以及在当 前帧的参考帧内确定与当前帧的 Tile相对应的参考帧的 Tile , 与当前帧的 Tile 相对应的参考帧的 Tile, 在参考帧中的位置与当前帧的 Tile在当前帧中的位置 相同, 并将确定的参考帧的 Tile的位置信息输出至预测单元 3。
预测单元 3 , 用于获取确定单元 2确定的参考帧的 Tile的位置信息, 在确 定的参考帧的 Tile内选择参考像素点,并根据参考像素点对所述 LCU内包含的
预测单元 PU进行预测, 将预测结果输出至编码单元 4。
编码单元 4, 用于获取预测单元 3输出的对 PU进行预测得到的预测结果, 并根据对 PU进行预测得到的预测结果, 对 LCU进行预测编码。
其中, 预测单元 3 , 具体用于:
根据待预测编码的 LCU在当前帧的 Tile中所处的位置,在确定的参考帧的 Tile内选取参考像素点, 其中, 选取的参考像素点所属的 LCU在参考帧的 Tile 中的位置, 与待预测编码的 LCU在当前帧的 Tile中的位置相同或相邻。 当前帧的参考帧中位置与当前帧的 Tile在当前帧的位置相同的参考帧的 Tile内 选取 PU的参考像素点, 使得当前预测编码的 LCU进行 PU预测时, 将参考帧 内对应当前帧的 Tile的参考帧的 Tile内的像素点作为参考像素点, 不会将参考 帧内其他 Tile内的像素点作为参考像素点, 能够防止某个 Tile发生的错误在整 帧图像内扩散, 以提高图像编码的精确性。 进一步的, 在选取参考像素点时, 根据待预测编码的 LCU所在当前帧的 Tile的位置在参考帧的 Tile内根据实际 情况进行参考像素点的选取, 能够进一步准确的预测得到待预测编码 LCU, 进 一步提高图像编码的精确性。
实施例七
基于与实施例三和实施例四涉及的图像预测编解码方法, 本发明实施例提 供了一种图像编码器, 该编码器包括用于接收待编码图像的接收单元 1 , 还包 括: 确定单元 2、 预测单元 3以及编码单元 4, 其中,
确定单元 2, 用于从接收单元 1获取接收到的待编码图像信息, 在当前帧 内确定待预测编码的最大编码单元 LCU所在的当前帧编码区块 Tile; 以及在当 前帧的参考帧内确定与 1当前帧的 Tile相对应的参考帧的 Tile, 与 1当前帧的 Tile相对应的参考帧的 Tile, 在参考帧中的位置与 1当前帧的 Tile在 1当前帧 中的位置相同, 并将确定的 1参考帧的 Tile的位置信息输出至 1预测单元 3。
预测单元 3 , 用于获取确定单元 2确定的参考帧的 Tile的位置信息, 并在 确定的参考帧的 Tile内选择时域候选运动矢量, 作为 LCU内包含的预测单元 PU的预测运动矢量 PMV, 根据 PMV预测得到 PU的运动矢量 MV, 并根据预 测得到的 MV对 PU进行预测, 将预测结果输出至编码单元 4。
编码单元 4, 用于获取预测单元 3输出的对 PU进行预测得到的预测结果, 并根据对 PU进行预测得到的预测结果, 对 LCU进行预测编码。
其中, 预测单元 3 , 具体用于:
在当前帧的参考帧内选择一个时域候选运动矢量;
判断选择的时域候选运动矢量是否在与当前帧的 Tile对应的参考帧的 Tile 内;
若是, 则利用选择的时域候选运动矢量, 作为 LCU内包含的 PU的 PMV; 若否,则返回执行继续在当前帧的参考帧内选择时域候选运动矢量的处理。 本发明实施例中, 在确定 PU的 PMV时, 在当前帧的参考帧内确定位置与 待预测编码的 LCU所在当前帧的 Tile在当前帧中的位置相同的 Tile作为参考 帧的 Tile,并在确定的参考帧的 Tile内选取时域候选运动矢量作为 PU的 PMV,
使得确定 PU的 PMV时 ,使用参考帧内对应待预测编码 LCU所在当前帧的 Tile 的参考帧的 Tile内的时域候选运动矢量, 不使用参考帧内其他 Tile的时域候选 运动矢量, 使得当某个 Tile发生错误时, 始终将错误限制在对应该 Tile的区域 内, 不会影响其他 Tile的预测编码, 从而使错误不会在整帧图像内扩散, 因此 提高了图像编码 ^精确性。 。 、 、 t 明的精神和范围。 这样, 倘若本发明的这些修改和变型属于本发明权利要求及 其等同技术的范围之内, 则本发明也意图包含这些改动和变型在内。
Claims
1、 一种图像预测编码方法, 其特征在于, 该方法包括:
在接收到的当前帧内,确定待预测编码的最大编码单元 LCU所在的当前帧 的编码区块 Tile;
在当前帧的参考帧内, 确定与所述当前帧的 Tile相对应的 Tile, 其中所述 参考帧的 Tile在参考帧中的位置与所述当前帧的 Tile在所述当前帧中的位置相 同;
在确定的所述参考帧的 Tile内选择参考像素点, 并根据所述参考像素点对 所述 LCU内包含的预测单元 PU进行预测;
根据对所述 PU进行预测得到的预测结果, 对所述 LCU进行预测编码。
2、 如权利要求 1所述的方法, 其特征在于, 在确定的所述参考帧的 Tile 内选择参考像素点, 包括:
根据所述待预测编码的 LCU在当前帧的 Tile中所处的位置,在确定的所述 参考帧的 Tile内选取参考像素点,其中,选取的所述参考像素点所属的 LCU在 参 "帧的 Tile中的位置, 与所述待预测编码的 LCU在当前帧的 Tile中的位置 相同或相邻。
3、 一种图像预测编码方法, 其特征在于, 该方法包括:
在接收到的当前帧内,确定待预测编码的最大编码单元 LCU所在的当前帧 的编码区块 Tile;
在当前帧的参考帧内, 确定与所述当前帧的 Tile相对应的参考帧的 Tile, 其中所述参考帧的 Tile在参考帧中的位置与所述当前帧的 Tile在所述当前帧中 的位置相同;
在确定的所述参考帧的 Tile内选择时域候选运动矢量,作为所述 LCU内包 含的预测单元 PU的预测运动矢量 PMV;
根据所述 PMV预测得到所述 PU的运动矢量 MV, 并根据预测得到的所述 MV对所述 PU进行预测;
根据对所述 PU进行预测得到的预测结果, 对所述 LCU进行预测编码。
4、 如权利要求 3所述的方法, 其特征在于, 在当前帧的参考帧内确定了所 述当前帧的 Tile相对应的参考帧的 Tile之后, 该方法包括:
在当前帧的参考帧内选择一个时域候选运动矢量;
判断选择的时域候选运动矢量是否在与所述当前帧的 Tile对应的所述参考 帧的 Tile内;
若是, 则利用选择的时域候选运动矢量, 作为所述 LCU内包含的 PU的 PMV;
若否,则返回执行继续在当前帧的参考帧内选择时域候选运动矢量的处理。
5、 一种图像编码器, 包括用于接收待编码图像的接收单元, 其特征在于, 还包括: 确定单元、 预测单元以及编码单元, 其中,
所述确定单元, 用于从所述接收单元获取接收到的待编码图像信息, 在当 前帧内确定待预测编码的最大编码单元 LCU所在的当前帧编码区块 Tile; 以及
在当前帧的参考帧内确定与所述当前帧的 Tile相对应的参考帧的 Tile, 与所述 当前帧的 Tile相对应的参考帧的 Tile, 在参考帧中的位置与所述当前帧的 Tile 在所述当前帧中的位置相同, 并将确定的所述参考帧的 Tile的位置信息输出至 所述预测单元;
所述预测单元, 用于获取所述确定单元确定的参考帧的 Tile的位置信息, 在确定的所述参考帧的 Tile内选择参考像素点, 并根据所述参考像素点对所述 LCU内包含的预测单元 PU进行预测, 将预测结果输出至所述编码单元; 预测结果, 并根据对所述 PU进行预测得到的预测结果, 对所述 LCU进行预测 编码。
6、 如权利要求 5所述的图像编码器, 其特征在于, 所述预测单元, 具体用 于:
根据所述待预测编码的 LCU在当前帧的 Tile中所处的位置,在确定的所述 参考帧的 Tile内选取参考像素点,其中,选取的所述参考像素点所属的 LCU在 参 帧的 Tile中的位置, 与所述待预测编码的 LCU在当前帧的 Tile中的位置 相同或相邻。
7、 一种图像编码器, 包括用于接收待编码图像的接收单元, 其特征在于, 还包括: 确定单元、 预测单元以及编码单元, 其中,
所述确定单元, 用于从所述接收单元获取接收到的待编码图像信息, 在当 前帧内确定待预测编码的最大编码单元 LCU所在的当前帧编码区块 Tile; 以及 在当前帧的参考帧内确定与所述当前帧的 Tile相对应的参考帧的 Tile, 与所述 当前帧的 Tile相对应的参考帧的 Tile, 在参考帧中的位置与所述当前帧的 Tile 在所述当前帧中的位置相同, 并将确定的所述参考帧的 Tile的位置信息输出至 所述预测单元;
所述预测单元, 用于获取所述确定单元确定的参考帧的 Tile的位置信息, 并在确定的所述参考帧的 Tile内选择时域候选运动矢量,作为所述 LCU内包含 的预测单元 PU的预测运动矢量 PMV, 根据所述 PMV预测得到所述 PU的运 动矢量 MV, 并根据预测得到的所述 MV对所述 PU进行预测, 将预测结果输 出至所 ¾码单 ; 、; ^ 、 、 、; 、 预测结果, 并根据对所述 PU进行预测得到的预测结果, 对所述 LCU进行预测 编码。
8、 如权利要求 7所述的图像编码器, 其特征在于, 所述预测单元, 具体用 于:
在当前帧的参考帧内选择一个时域候选运动矢量;
判断选择的时域候选运动矢量是否在与所述当前帧的 Tile对应的所述参考 帧的 Tile内;
若是, 则利用选择的时域候选运动矢量, 作为所述 LCU内包含的 PU的 PMV;
若否,则返回执行继续在当前帧的参考帧内选择时域候选运动矢量的处理。
Priority Applications (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
EP13883015.3A EP2983365A4 (en) | 2013-04-26 | 2013-10-18 | IMAGE PREDICTIVE ENCODING METHOD AND IMAGE ENCODER |
US14/921,548 US20160044329A1 (en) | 2013-04-26 | 2015-10-23 | Image Predictive Coding Method and Image Encoder |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201310149199.X | 2013-04-26 | ||
CN201310149199.XA CN104125463A (zh) | 2013-04-26 | 2013-04-26 | 一种图像预测编码方法及图像编码器 |
Related Child Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US14/921,548 Continuation US20160044329A1 (en) | 2013-04-26 | 2015-10-23 | Image Predictive Coding Method and Image Encoder |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2014173092A1 true WO2014173092A1 (zh) | 2014-10-30 |
Family
ID=51770709
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/CN2013/085426 WO2014173092A1 (zh) | 2013-04-26 | 2013-10-18 | 一种图像预测编码方法及图像编码器 |
Country Status (4)
Country | Link |
---|---|
US (1) | US20160044329A1 (zh) |
EP (1) | EP2983365A4 (zh) |
CN (1) | CN104125463A (zh) |
WO (1) | WO2014173092A1 (zh) |
Families Citing this family (19)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US8942473B2 (en) * | 2012-07-25 | 2015-01-27 | Ko Hung Lin | Image processing method and display apparatus |
CN105721866B (zh) * | 2016-02-01 | 2019-02-12 | 北京奇艺世纪科技有限公司 | 一种编码单元划分方法及装置 |
CN107135397B (zh) * | 2017-04-28 | 2018-07-06 | 中国科学技术大学 | 一种全景视频编码方法和装置 |
CN110662056B (zh) | 2018-06-29 | 2022-06-07 | 北京字节跳动网络技术有限公司 | 哪个查找表需要更新或不更新 |
JP7460617B2 (ja) | 2018-06-29 | 2024-04-02 | 北京字節跳動網絡技術有限公司 | Lut更新条件 |
EP3791588A1 (en) | 2018-06-29 | 2021-03-17 | Beijing Bytedance Network Technology Co. Ltd. | Checking order of motion candidates in lut |
WO2020003278A1 (en) | 2018-06-29 | 2020-01-02 | Beijing Bytedance Network Technology Co., Ltd. | Update of look up table: fifo, constrained fifo |
CN110662057B (zh) | 2018-06-29 | 2022-06-21 | 北京字节跳动网络技术有限公司 | 视频处理方法、装置、设备以及存储比特流的方法 |
EP3797516A1 (en) | 2018-06-29 | 2021-03-31 | Beijing Bytedance Network Technology Co. Ltd. | Interaction between lut and amvp |
KR20210024502A (ko) | 2018-06-29 | 2021-03-05 | 베이징 바이트댄스 네트워크 테크놀로지 컴퍼니, 리미티드 | Hmvp 후보를 병합/amvp에 추가할 때의 부분/풀 프루닝 |
TWI723444B (zh) | 2018-06-29 | 2021-04-01 | 大陸商北京字節跳動網絡技術有限公司 | 使用一個或多個查找表來按順序存儲先前編碼的運動信息並使用它們來編碼後面的塊的概念 |
CN110662063B (zh) * | 2018-06-29 | 2020-12-15 | 北京字节跳动网络技术有限公司 | 视频处理方法、装置和计算机可读存储介质 |
EP4307679A3 (en) | 2018-07-02 | 2024-06-19 | Beijing Bytedance Network Technology Co., Ltd. | Luts with intra prediction modes and intra mode prediction from non-adjacent blocks |
TW202025760A (zh) | 2018-09-12 | 2020-07-01 | 大陸商北京字節跳動網絡技術有限公司 | 要檢查多少個hmvp候選 |
CN116456082A (zh) | 2018-12-25 | 2023-07-18 | Oppo广东移动通信有限公司 | 编码预测方法、装置及计算机存储介质 |
CN113273186A (zh) | 2019-01-10 | 2021-08-17 | 北京字节跳动网络技术有限公司 | Lut更新的调用 |
CN113383554B (zh) | 2019-01-13 | 2022-12-16 | 北京字节跳动网络技术有限公司 | LUT和共享Merge列表之间的交互 |
CN113302937B (zh) | 2019-01-16 | 2024-08-02 | 北京字节跳动网络技术有限公司 | 运动候选推导 |
CN113615193B (zh) | 2019-03-22 | 2024-06-25 | 北京字节跳动网络技术有限公司 | Merge列表构建和其他工具之间的交互 |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1756364A (zh) * | 2004-09-30 | 2006-04-05 | 华为技术有限公司 | 帧内预测模式的选择方法 |
JP3833585B2 (ja) * | 2002-07-24 | 2006-10-11 | Kddi株式会社 | 画像符号化装置および画像符号化方法、コンピュータプログラム |
CN101605256A (zh) * | 2008-06-12 | 2009-12-16 | 华为技术有限公司 | 一种视频编解码的方法及装置 |
CN102611885A (zh) * | 2011-01-20 | 2012-07-25 | 华为技术有限公司 | 一种编解码方法和装置 |
WO2012098845A1 (ja) * | 2011-01-18 | 2012-07-26 | 株式会社日立製作所 | 画像符号化方法、画像符号化装置、画像復号方法及び画像復号装置 |
Family Cites Families (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US9008170B2 (en) * | 2011-05-10 | 2015-04-14 | Qualcomm Incorporated | Offset type and coefficients signaling method for sample adaptive offset |
CN102883163B (zh) * | 2012-10-08 | 2014-05-28 | 华为技术有限公司 | 用于运动矢量预测的运动矢量列表建立的方法、装置 |
CN103067714B (zh) * | 2013-01-09 | 2016-08-03 | 北京大学 | 一种运动矢量预测的方法 |
US9749627B2 (en) * | 2013-04-08 | 2017-08-29 | Microsoft Technology Licensing, Llc | Control data for motion-constrained tile set |
-
2013
- 2013-04-26 CN CN201310149199.XA patent/CN104125463A/zh active Pending
- 2013-10-18 EP EP13883015.3A patent/EP2983365A4/en not_active Withdrawn
- 2013-10-18 WO PCT/CN2013/085426 patent/WO2014173092A1/zh active Application Filing
-
2015
- 2015-10-23 US US14/921,548 patent/US20160044329A1/en not_active Abandoned
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP3833585B2 (ja) * | 2002-07-24 | 2006-10-11 | Kddi株式会社 | 画像符号化装置および画像符号化方法、コンピュータプログラム |
CN1756364A (zh) * | 2004-09-30 | 2006-04-05 | 华为技术有限公司 | 帧内预测模式的选择方法 |
CN101605256A (zh) * | 2008-06-12 | 2009-12-16 | 华为技术有限公司 | 一种视频编解码的方法及装置 |
WO2012098845A1 (ja) * | 2011-01-18 | 2012-07-26 | 株式会社日立製作所 | 画像符号化方法、画像符号化装置、画像復号方法及び画像復号装置 |
CN102611885A (zh) * | 2011-01-20 | 2012-07-25 | 华为技术有限公司 | 一种编解码方法和装置 |
Non-Patent Citations (1)
Title |
---|
See also references of EP2983365A4 * |
Also Published As
Publication number | Publication date |
---|---|
US20160044329A1 (en) | 2016-02-11 |
EP2983365A1 (en) | 2016-02-10 |
CN104125463A (zh) | 2014-10-29 |
EP2983365A4 (en) | 2016-03-30 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
WO2014173092A1 (zh) | 一种图像预测编码方法及图像编码器 | |
CN113170099B (zh) | 块内拷贝模式和帧间预测工具之间的交互 | |
CN112889269B (zh) | 视频解码方法及装置 | |
CN110719469B (zh) | 视频编解码方法、设备和存储介质 | |
TWI619374B (zh) | 利用仿射運動補償的視訊編碼方法與裝置 | |
JP2021002892A (ja) | 適応的映像符号化装置及び方法 | |
CN113141783B (zh) | 一种视频编解码方法及电子装置 | |
CN101179728B (zh) | 帧间编码模式确定方法及装置 | |
CN111373749B (zh) | 视频编码和解码中的低复杂度双向帧内预测的方法和装置 | |
JP5593440B2 (ja) | 動きベクトル予測方法 | |
TW201424399A (zh) | 視訊編碼及解碼方法及使用該方法之裝置 | |
JP2023052767A (ja) | 映像処理方法及び符号化器 | |
CN113055682B (zh) | 一种编解码方法、装置及其设备 | |
KR20100039247A (ko) | 비디오 신호 처리 방법 및 장치 | |
CN110351552B (zh) | 视频编码中一种快速编码方法 | |
JP5938424B2 (ja) | 画像ブロックを再構成および符号化する方法 | |
CN106331699B (zh) | 图像编码方法和图像编码器 | |
RU2810318C1 (ru) | Способ и устройство для кодирования видео | |
CN103379322A (zh) | 一种先进运动矢量预测amvp的并行实现方法、装置及系统 | |
JP7523844B2 (ja) | デブロッキング操作の高度制御 | |
WO2024000581A1 (zh) | 排序方法、解码方法、编码方法、装置、解码器和编码器 | |
WO2021056219A1 (zh) | 视频编解码的方法和装置 | |
CN118042136A (zh) | 编解码方法和装置 | |
TW202145783A (zh) | 幀間預測方法、編碼器、解碼器以及電腦儲存媒介 | |
CN117397238A (zh) | 编解码方法和装置 |
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: 13883015 Country of ref document: EP Kind code of ref document: A1 |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2013883015 Country of ref document: EP |