Encoding of the Link to a Reference Block in Video Compression by Image Content Based on Search and Ranking
Method and apparatus are provided for efficient reference data encoding for video compression by image content based search and ranking. An apparatus includes a rank transformer for respectively transforming reference data for each of a plurality of candidate reference blocks with respect to a current block to be encoded into a respective rank number there for based on a context feature of the current block with respect to the context feature of each of the plurality of candidate reference blocks. The apparatus further includes an entropy encoder for respectively entropy encoding the respective rank number for each of the plurality of candidate reference blocks with respect to the current block in place of, and representative of, the reference data for each of the plurality of candidate reference blocks with respect to the current block.
1 . An apparatus, comprising:
a rank transformer for respectively transforming reference data for each of a plurality of candidate reference blocks with respect to a current block to be encoded into a respective rank number there for based on a context feature of said current block with respect to said context feature of each of said plurality of candidate reference blocks, and for determining a rank number, said rank number corresponding to a particular one of said plurality of candidate reference blocks to be actually used as a reference block to encode said current block; and
an entropy encoder ( 220 ) for entropy encoding said determined rank number said current block.
2 . The apparatus of claim 1 , wherein said reference data comprises at least one of motion vectors, spatial displacement vectors, and disparity values.
3 . The apparatus of claim 1 , wherein said reference data for each of said plurality of candidate reference blocks with respect to said current block is respectively transformed into said respective rank number there for by calculating a respective distance between said current block and a respective one of said plurality of candidate reference blocks based on said context feature of said current block with respect to said context feature of said respective one of said plurality of candidate blocks, repeating said calculating of said respective distance for remaining ones of said plurality of candidate reference blocks to obtain a plurality of respective distances, and sorting said plurality of candidate reference blocks based on said plurality of respective distances to obtain a rank list having said plurality of respective rank numbers specified therein.
4 . The apparatus of claim 1 , wherein said particular one of said plurality of candidate reference blocks to be actually used as said reference block to encode said current block is identified based on a respective one of said rank numbers corresponding thereto indicating a closest match between said particular one of said plurality of candidate reference blocks and said current block based on a shortest one of said plurality of respective distances.
5 . The apparatus of claim 1 , wherein said context feature for said current block comprises pixels surrounding said current block, and said context feature for each of said plurality of candidate reference blocks respectively comprises pixels respectively surrounding there around.
6 . The apparatus of claim 1 , wherein said context feature of each of said plurality of candidate reference blocks form a searchable context feature set that is compared to said context feature of said current block.
7 . The apparatus of claim 1 , wherein said context feature of said current block is matched to said context feature of each of said plurality of candidate reference blocks based on respective distances there between.
8 . A method, comprising:
respectively transforming reference data for each of a plurality of candidate reference blocks with respect to a current block to be encoded into a respective rank number there for based on a context feature of said current block with respect to said context feature of each of said plurality of candidate reference blocks;
determining a rank number for said current block, said rank number corresponding to a particular one of said plurality of candidate reference blocks to be actually used as a reference block to encode said current block; and
entropy encoding said determined rank number for said current block.
9 . The method of claim 8 , wherein said reference data comprises at least one of motion vectors, spatial displacement vectors, and disparity values.
10 . The method of claim 8 , wherein said reference data for each of said plurality of candidate reference blocks with respect to said current block is respectively transformed into said respective rank number there for by calculating a respective distance between said current block and a respective one of said plurality of candidate reference blocks based on said context feature of said current block with respect to said context feature of said respective one of said plurality of candidate blocks, repeating said calculating of said respective distance for remaining ones of said plurality of candidate reference blocks to obtain a plurality of respective distances, and sorting said plurality of candidate reference blocks based on said plurality of respective distances to obtain a rank list having said plurality of respective rank numbers specified therein.
11 . The method of claim 10 , wherein a particular one of said plurality of candidate reference blocks to be actually used as a reference block to encode said current block is identified based on a respective one of said rank numbers corresponding thereto indicating a closest match between said particular one of said plurality of candidate reference blocks and said current block based on a shortest one of said plurality of respective distances.
12 . The method of claim 8 , wherein said context feature for said current block comprises pixels surrounding said current block, and said context feature for each of said plurality of candidate reference blocks respectively comprises pixels respectively surrounding there around.
13 . The method of claim 8 , wherein said context feature of each of said plurality of candidate reference blocks form a searchable context feature set that is compared to said context feature of said current block.
14 . The method of claim 8 , wherein said context feature of said current block is matched to said context feature of each of said plurality of candidate reference blocks based on respective distances there between.
15 . An apparatus, comprising:
means for respectively transforming reference data for each of a plurality of candidate reference blocks with respect to a current block to be encoded into a respective rank number there for based on a context feature of said current block with respect to said context feature of each of said plurality of candidate reference blocks, said transforming means determining a rank number for said current block, said rank number corresponding to a particular one of said plurality of candidate reference blocks to be actually used as a reference block to encode said current block; and
means ( 220 ) for entropy encoding said determined rank number for said current block.
16 . The apparatus of claim 15 , wherein said reference data comprises at least one of motion vectors, spatial displacement vectors, and disparity values.
17 . The apparatus of claim 15 , wherein said reference data for each of said plurality of candidate reference blocks with respect to said current block is respectively transformed into said respective rank number there for by calculating a respective distance between said current block and a respective one of said plurality of candidate reference blocks based on said context feature of said current block with respect to said context feature of said respective one of said plurality of candidate blocks, repeating said calculating of said respective distance for remaining ones of said plurality of candidate reference blocks to obtain a plurality of respective distances, and sorting said plurality of candidate reference blocks based on said plurality of respective distances to obtain a rank list having said plurality of respective rank numbers specified therein.
18 . The apparatus of claim 16 , wherein said particular one of said plurality of candidate reference blocks to be actually used as said reference block to encode said current block is identified based on a respective one of said rank numbers corresponding thereto indicating a closest match between said particular one of said plurality of candidate reference blocks and said current block based on a shortest one of said plurality of respective distances.
19 . The apparatus of claim 15 , wherein said context feature for said current block comprises pixels surrounding said current block, and said context feature for each of said plurality of candidate reference blocks respectively comprises pixels respectively surrounding there around.
20 . The apparatus of claim 15 , wherein said context feature of each of said plurality of candidate reference blocks form a searchable context feature set that is compared to said context feature of said current block.
21 . The apparatus of claim 15 , wherein said context feature of said current block is matched to said context feature of each of said plurality of candidate reference blocks based on respective distances there between.