Devices and methods for context reduction in last significant coefficient position coding
View Patent ↗In one embodiment, a method of video encoding, the method comprising: encoding a position of a last non-zero coefficient within a video block, wherein the position of the last non-zero coefficient is provided in a significance map according to block width; and assigning at least one context model to the significance map; wherein the at least one context model includes at least three first contexts, and wherein each of the first three contexts is configured to be shared across two or more blocks of different widths.
1. A method of video coding, the method comprising:
generating a context model comprising at least a first context, a second context, and a third context, wherein the context model comprises:
the first context assigned to a first significant bin for a 4×4 block, a first and second significant bins for an 8×8 block, and a first, second, third and fourth most significant bins for a 16×16 block,
the second context assigned to a second significant bin for the 4×4 block, a third and fourth significant bins for the 8×8 block, and a fifth, sixth and seventh significant bins for the 16×16 block, and
the third context assigned to a third significant bin for the 4×4 block and a fifth significant bin for the 8×8 block;
entropy coding, using the context model, a position of a last non-zero coefficient of a first video block having a size corresponding to one of the 4×4 block, the 8×8 block, or the 16×16 block; and
entropy coding transform coefficients of the video block.
2. The method of claim 1 , further comprising:
generating a significance map for a second block comprising a plurality of non-overlapping sub-blocks, the first video block being one of the plurality of non-overlapping sub-blocks, the significance map identifying which of the plurality of non-overlapping sub-blocks comprises at least one non-zero transform coefficient and which of the plurality of non-overlapping sub-blocks comprises only zero transform coefficients; and
entropy coding a position of a last block of the plurality of non-overlapping sub-blocks having only zero transform coefficients within the significance map.
3. The method of claim 1 , wherein entropy coding the position of the last non-zero coefficient comprises context adaptive binary arithmetic coding (CABAC).
4. The method of claim 3 , wherein CABAC includes truncated codewords for the position of the last non-zero coefficient.
5. The method of claim 1 , wherein entropy coding the position of the last non-zero coefficient comprises signaling an X coordinate and a Y coordinate the position of the last non-zero coefficient independently.
6. The method of claim 1 , wherein entropy coding the transform coefficients occurs after entropy coding the position of the last non-zero transform coefficient.
7. The method of claim 1 , wherein entropy coding the transform coefficients comprises high efficiency video coding (HEVC).
8. The method of claim 1 , wherein entropy coding the transform coefficients comprises entropy decoding.
9. The method of claim 1 , wherein entropy coding the transform coefficients comprises entropy encoding.
10. The method of claim 1 , wherein the method is implemented on a computer having a processor and a memory coupled to said processor, wherein at least some of steps are performed using said processor.
11. An apparatus for coding video, the apparatus comprising a video coder configured to:
generate a context model comprising at least a first context, a second context, and a third context, wherein the context model comprises:
the first context assigned to a first significant bin for a 4×4 block, a first and second significant bins for an 8×8 block, and a first, second, third and fourth most significant bins for a 16×16 block,
the second context assigned to a second significant bin for the 4×4 block, a third and fourth significant bins for the 8×8 block, and a fifth, sixth and seventh significant bins for the 16×16 block, and
the third context assigned to a third significant bin for the 4×4 block and a fifth significant bin for the 8×8 block;
entropy code, using the context model, a position of a last non-zero coefficient of a first video block having a size corresponding to one of the 4×4 block, the 8×8 block, or the 16×16 block; and
entropy code transform coefficients of the video block.
12. The apparatus of claim 11 , wherein the video coder comprises a video encoder including at least one of:
an integrated circuit; or
a microprocessor.
13. The apparatus of claim 11 , wherein the video coder comprises a video decoder including at least one of:
an integrated circuit; or
a microprocessor.
14. The apparatus of claim 11 , wherein the video coder is configured to:
generate a significance map for a second block comprising a plurality of non-overlapping sub-blocks, the first video block being one of the plurality of non-overlapping sub-blocks, the significance map identifying which of the plurality of non-overlapping sub-blocks comprises at least one non-zero transform coefficient and which of the plurality of non-overlapping sub-blocks comprises only zero transform coefficients; and
entropy code a position of a last block of the plurality of non-overlapping sub-blocks having only zero transform coefficients within the significance map.
15. The apparatus of claim 11 , wherein entropy coding the position of the last non-zero coefficient comprises context adaptive binary arithmetic coding (CABAC) including truncated codewords for the position of the last non-zero coefficient.
16. The apparatus of claim 11 , wherein the video coder is configured to entropy code the transform coefficients after entropy coding the position of the last non-zero transform coefficient.