Method and device for encoding/decoding image using feature vectors of surrounding blocks
A video encoding apparatus, a decoding apparatus and method thereof by using feature vectors of at least one neighboring block are provided. In the video encoding apparatus, the decoding apparatus and the method, the feature vector is extracted from pixels of at least one neighboring block, and an optimal prediction mode represented by the feature vector is extracted with reference to a preset feature vector distribution of image blocks. The extracted optimal prediction mode is decided as a most probable mode (MPM) or an intra mode of a current block, and the current block is intra-predicted and encoded or decoded by using pixels of the at least one neighboring block adjacent to the current block.
1. A video encoding apparatus, comprising:
an intra prediction encoder configured to determine an intra mode of a current block from among a plurality of intra modes, and
perform an intra prediction and encoding on the current block based on the intra mode of the current block by using pixels of at least one neighboring block adjacent to the current block;
a feature vector extractor configured to generate a feature vector from pixels of the at least one neighboring block, wherein the feature vector is generated from a plurality of differences between values of adjacent pixels of the at least one neighboring block, respective differences of the plurality of differences are calculated based on values of the adjacent pixels located in respective directions of a plurality of directions, the feature vector includes a plurality of elements, and respective elements of the plurality of elements correspond to respective differences of the plurality of differences; and
a prediction mode encoder configured to extract, from a preset feature vector distribution, a prediction mode identified by the feature vector, the preset feature vector distribution including respective feature vectors for respective intra modes, and
encode the identified prediction mode as a most probable mode (MPM) for the intra mode of the current block,
wherein the MPM is used for predicting the intra mode of the current block, and a length of encoded data for indicating the intra mode of the current block is determined according to whether the intra mode of the current block is equal to the MPM.
2. The video encoding apparatus of claim 1 , wherein the plurality of directions are determined from directions including a vertical direction, a horizontal direction, a diagonal down-right direction, a diagonal down-left direction, a horizontal-down (width/height ratio=1/2) direction, a horizontal-up (width/height ratio=1/2) direction, a horizontal-down (width/height ratio=2) direction, and a horizontal-up (width/height ratio=2) direction.
3. A video decoding apparatus, comprising:
a feature vector extractor configured to generate a feature vector from pixels of at least one neighboring block adjacent to a current block, wherein the feature vector is generated from a plurality of differences between values of adjacent pixels of the at least one neighboring block, respective differences of the plurality of differences are calculated based on values of the adjacent pixels located in respective directions of a plurality of directions, the feature vector includes a plurality of elements, and respective elements of the plurality of elements correspond to respective differences of the plurality of differences;
a prediction mode processor configured to extract, from a preset feature vector distribution, a prediction mode identified by the feature vector, the preset feature vector distribution including respective feature vectors for respective intra modes,
decide the extracted prediction mode as a most probable mode (MPM) for an intra mode of the current block, wherein the MPM is used for predicting the intra mode of the current block, and a length of encoded data for indicating the intra mode of the current block is determined according to whether the intra mode of the current block is equal to the MPM; and
an intra prediction decoder configured to generate a predicted block by performing an intra prediction based on the intra mode by using pixels of the at least one neighboring block to reconstruct the current block.
4. The video decoding apparatus of claim 3 , wherein the plurality of directions are determined from directions including a vertical direction, a horizontal direction, a diagonal down-right direction, a diagonal down-left direction, a horizontal-down (width/height ratio=1/2) direction, a horizontal-up (width/height ratio=1/2) direction, a horizontal-down (width/height ratio=2) direction, and a horizontal-up (width/height ratio=2) direction.
5. A video encoding method performed by a video encoding apparatus, comprising:
determining an intra mode of a current block from among a plurality of intra modes;
performing an intra prediction and encoding on the current block based on the intra mode of the current block by using pixels of at least one neighboring block adjacent to the current block;
generating a feature vector from pixels of the at least one neighboring block, wherein the feature vector is generated from a plurality of differences between values of adjacent pixels of the at least one neighboring block, respective differences of the plurality of differences are calculated based on values of the adjacent pixels located in respective directions of a plurality of directions, the feature vector includes a plurality of elements, and respective elements of the plurality of elements correspond to respective differences of the plurality of differences;
extracting, from a preset feature vector distribution, a prediction mode identified by the feature vector, the preset feature vector distribution including respective feature vectors for respective intra modes; and
encoding the identified prediction mode as a most probable mode (MPM) for the intra mode of the current block,
wherein the MPM is used for predicting the intra mode of the current block, and a length of encoded data for indicating the intra mode of the current block is determined according to whether the intra mode of the current block is equal to the MPM.
6. The video encoding method of claim 5 , wherein the plurality of directions are determined from directions including a vertical direction, a horizontal direction, a diagonal down-right direction, a diagonal down-left direction, a horizontal-down (width/height ratio=1/2) direction, a horizontal-up (width/height ratio=1/2) direction, a horizontal-down (width/height ratio=2) direction, and a horizontal-up (width/height ratio=2) direction.
7. A video decoding method performed by a video decoding apparatus, comprising:
generating a feature vector from pixels of at least one neighboring block adjacent to a current block, wherein the feature vector is generated from a plurality of differences between values of adjacent pixels of the at least one neighboring block, respective differences of the plurality of differences are calculated based on values of the adjacent pixels located in respective directions of a plurality of directions, the feature vector including a plurality of elements, and respective elements of the plurality of elements corresponding to respective differences of the plurality of difference;
extracting, from a preset feature vector distribution, a prediction mode identified by the feature vector, the preset feature vector distribution including respective feature vectors for respective intra modes;
deciding the extracted prediction mode as a most probable mode (MPM) for an intra mode of the current block, wherein the MPM is used for predicting the intra mode of the current block, and a length of encoded data for indicating the intra mode of the current block is determined according to whether the intra mode of the current block is equal to the MPM; and
generating a predicted block of the current block by performing an intra prediction based on the intra mode by using pixels of the at least one neighboring block to thereby reconstruct the current block.
8. The video decoding method of claim 7 , wherein the plurality of directions are determined from directions including a vertical direction, a horizontal direction, a diagonal down-right direction, a diagonal down-left direction, a horizontal-down (width/height ratio=1/2) direction, a horizontal-up (width/height ratio=1/2) direction, a horizontal-down (width/height ratio=2) direction, and a horizontal-up (width/height ratio=2) direction.