Karhunen loeve transform for video coding
In one implementation, KLT transform matrices are derived during the encoding or decoding process. In particular, the KLT transform matrices can be derived for either the primary transforming stage or the secondary transforming stage, or both. Because the KLT transform matrices can be derived at both the encoder and decoder sides, the KLT transform matrices do not need to be signaled in the bitstream. To derive the KLT transform matrix for a current block to be encoded or decoded, a template for the current block is defined, and is used to search in reconstructed regions multiple blocks with templates similar to the current template. Those multiple blocks are used to train the KLT transform matrix.
1. A method for video encoding, comprising:
obtaining a prediction block and a residual block for a current block;
obtaining a plurality of residual blocks, based on a plurality of respective neighboring blocks and said prediction block for said current block;
applying a primary transform to said plurality of residual blocks to obtain a plurality of respective transformed residual blocks;
obtaining a Karhunen Loeve Transform (KLT) transform matrix based on said plurality of transformed residual blocks;
applying said primary transform to said residual block for said current block, to obtain a current transformed residual block; and
applying said KLT transform matrix as secondary transform to said current transformed residual block.
2. The method of claim 1 , wherein said primary transform is Discrete Cosine Transform type II (DCT-II).
3. The method of claim 1 , wherein said primary transform is another KLT transform.
4. The method of claim 1 , further comprising:
encoding a signal indicating that said KLT transform matrix is to be used to transform said residual block for said current block; and
disabling one or more other transform types for said current block.
5. The method of claim 4 , wherein signaling indicating that said KLT transform matrix is to be used is independent from signaling of a prediction type of said current block.
6. An apparatus for video encoding, comprising:
at least a memory and one or more processors, wherein said one or more processors are configured to:
obtain a prediction block and a residual block for a current block;
obtain a plurality of residual blocks, based on a plurality of respective neighboring blocks and said prediction block for said current block;
apply a primary transform to said plurality of residual blocks to obtain a plurality of respective transformed residual blocks;
obtain a Karhunen Loeve Transform (KLT) transform matrix based on said plurality of transformed residual blocks;
apply said primary transform to said residual block for said current block, to obtain a current transformed residual block; and
apply said KLT transform matrix as secondary transform to said current transformed residual block.
7. The apparatus of claim 6 , wherein said primary transform is Discrete Cosine Transform type II (DCT-II).
8. The apparatus of claim 6 , wherein said primary transform is another KLT transform.
9. The apparatus of claim 6 , wherein said one or more processors are further configured to:
encode a signal indicating that said KLT transform matrix is to be used to transform said residual block for said current block; and
disable one or more other transform types for said current block.
10. The apparatus of claim 9 , wherein signaling indicating that said KLT transform matrix is to be used is independent from signaling of a prediction type of said current block.
11. A method for video decoding, comprising:
obtaining a prediction block and transform coefficients for a current block;
obtaining a plurality of residual blocks, based on a plurality of respective neighboring blocks and said prediction block for said current block;
applying a primary transform to said plurality of residual blocks to obtain a plurality of respective transformed residual blocks;
obtaining a Karhunen Loeve Transform (KLT) transform matrix based on said plurality of transformed residual blocks;
applying said KLT transform matrix as inverse secondary transform to said transform coefficients of said current block to obtain a current inverse secondary transformed block; and
applying an inverse of said primary transform to said current inverse secondary transformed block to reconstruct a residual block for said current block.
12. The method for claim 11 , further comprising:
decoding a signal indicating that said KLT transform matrix is to be used for inverse transform to reconstruct said residual block for said current block; and
inferring that one or more other transform types are disabled for said current block.
13. The method of claim 12 , wherein said primary transform is Discrete Cosine Transform type II (DCT-II).
14. The method of claim 12 , wherein said primary transform is another KLT transform.
15. The method of claim 12 , further comprising:
decoding a signal indicating that said KLT transform matrix is to be used to inverse transform to reconstruct said residual block for said current block.
16. An apparatus for video decoding, comprising:
at least a memory and one or more processors, wherein said one or more processors are configured to:
obtain a prediction block and transform coefficients for a current block;
obtain a plurality of residual blocks, based on a plurality of respective neighboring blocks and said prediction block for said current block;
apply a primary transform to said plurality of residual blocks to obtain a plurality of respective transformed residual blocks;
obtain a Karhunen Loeve Transform (KLT) transform matrix based on said plurality of transformed residual blocks;
apply said KLT transform matrix as inverse secondary transform to said transform coefficients of said current block to obtain a current inverse secondary transformed block; and
apply an inverse of said primary transform to said current inverse secondary transformed block to reconstruct a residual block for said current block.
17. The apparatus of claim 16 , wherein said one or more processors are configured to:
decode a signal indicating that said KLT transform matrix is to be used for inverse transform to reconstruct said residual block for said current block; and
infer that one or more other transform types are disabled for said current block.
18. The apparatus of claim 16 , wherein said primary transform is Discrete Cosine Transform type II (DCT-II).
19. The apparatus of claim 18 , wherein said primary transform is another KLT transform.
20. The apparatus of claim 18 , wherein said one or more processors are further configured to:
decode a signal indicating that said KLT transform matrix is to be used for inverse transform to reconstruct said residual block for said current block; and
infer that one or more other transform types are disabled for said current block.