Adaptive loop filter with enhanced classification methods
An apparatus configured to decode video data, including a memory configured to store a block of a current picture of the video data, and one or more processors in communication with the memory. The one or more processors may be configured to reconstruct the block of the current picture of the video data, determine a classification, from among a plurality of classifications, of the block according to a rule, wherein the rule is adaptively determined based on characteristics of coded sequences of the video data, and apply an adaptive loop filter (ALF) to the block based on the classification of the block.
1. An apparatus configured to decode video data, the apparatus comprising:
a memory configured to store a block of a current picture of the video data; and
one or more processors in communication with the memory, the one or more processors configured to:
reconstruct the block of the current picture of the video data;
determine a classification, from among a plurality of classifications, of the block according to a rule, wherein the rule is adaptively determined based on characteristics of coded sequences of the video data, and wherein the rule is based on a bit-depth of a pixel in the block; and
apply an adaptive loop filter (ALF) to the block based on the classification of the block.
2. The apparatus of claim 1 , wherein, to determine the classification, the one or more processors may be further configured to:
adjust a quantization process for determining an activity value of the block of the current picture based on the rule.
3. The apparatus of claim 2 , wherein, to adjust the quantization process, the one or more processors may be configured to:
determine a function of mapping the activity value of the block of the current picture to an activity index based on the rule; and
determine the classification based on the activity index.
4. The apparatus of claim 1 , wherein, to determine the classification, the one or more processors may be further configured to:
determine one or more thresholds for determining a direction value of the block of the current picture based on the rule;
determine the direction value of the block of the current picture based on the one or more thresholds; and
determine the classification based on the direction value.
5. The apparatus of claim 1 , wherein the ALF is a geometric transformation-based ALF (GALF).
6. The apparatus of claim 1 , wherein the one or more processors are further configured to:
obtain, from a bitstream that comprises an encoded representation of the current picture, an indication of the rule.
7. The apparatus of claim 1 , wherein the one or more processors are further configured to:
derive the rule based on characteristics of coded sequences of the video data.
8. The apparatus of claim 1 , wherein the rule indicates how to merge blocks with different activity values into one class.
9. The apparatus of claim 8 , wherein the one or more processors are further configured to obtain, from a bitstream that comprises an encoded representation of the current picture, an indication of the rule that indicates how to merge blocks with different activity values into one class.
10. The apparatus of claim 1 , wherein the rule indicates how to merge blocks with different direction values into one class.
11. The apparatus of claim 10 , wherein the one or more processors are further configured to obtain, from a bitstream that comprises an encoded representation of the current picture, an indication of the rule that indicates how to merge blocks with different direction values into one class.
12. The apparatus of claim 1 , wherein to determine the classification, the one or more processors are further configured to:
determine the classification, from among the plurality of classifications, of the block according to the rule, wherein the rule indicates how to merge blocks with different activity values and direction values into one class.
13. The apparatus of claim 1 , wherein the one or more processors are further configured to adaptively change a total number of allowed classifications of the plurality of classifications, per unit, wherein the unit is one of a tile, slice, picture, sequence, or largest coding unit (LCU).
14. The apparatus of claim 1 , wherein the one or more processors are further configured to:
output the block after applying the ALF.
15. A method for decoding video data, the method comprising:
reconstructing a block of a current picture of the video data;
determining a classification, from among a plurality of classifications, of the block according to a rule, wherein the rule is adaptively determined based on characteristics of coded sequences of the video data, and wherein the rule is based on a bit-depth of a pixel in the block; and
applying an adaptive loop filter (ALF) to the block based on the classification of the block.
16. The method of claim 15 , wherein determining the classification comprises:
adjusting a quantization process for determining an activity value of the block of the current picture based on the rule.
17. The method of claim 16 , wherein adjusting the quantization process comprises:
determining a function of mapping the activity value of the block of the current picture to an activity index based on the rule; and
determining the classification based on the activity index.
18. The method of claim 15 , wherein determining the classification comprises:
determining one or more thresholds for determining a direction value of the block of the current picture based on the rule;
determining the direction value of the block of the current picture based on the one or more thresholds; and
determining the classification based on the direction value.
19. The method of claim 15 , wherein the ALF is a geometric transformation-based ALF (GALF).
20. The method of claim 15 , further comprising:
obtaining, from a bitstream that comprises an encoded representation of the current picture, an indication of the rule.
21. The method of claim 15 , further comprising:
deriving the rule based on characteristics of coded sequences of the video data.
22. The method of claim 15 , wherein the rule indicates how to merge blocks with different activity values into one class.
23. The method of claim 22 , further comprising:
obtaining, from a bitstream that comprises an encoded representation of the current picture, an indication of the rule that indicates how to merge blocks with different activity values into one class.
24. The method of claim 15 , wherein the rule indicates how to merge blocks with different direction values into one class.
25. The method of claim 24 , further comprising:
obtaining, from a bitstream that comprises an encoded representation of the current picture, an indication of the rule that indicates how to merge blocks with different direction values into one class.
26. The method of claim 15 , further comprising:
determining the classification, from among the plurality of classifications, of the block according to the rule, wherein the rule indicates how to merge blocks with different activity values and direction values into one class.
27. The method of claim 15 , further comprising:
adaptively changing a total number of allowed classifications of the plurality of classifications, per unit, wherein the unit is one of a tile, slice, picture, sequence, or largest coding unit (LCU).
28. The method of claim 15 , further comprising:
outputting the block after applying the ALF.
29. A non-transitory computer-readable storage medium storing instructions that, when executed, causes one or more processors of a device configured to decode video data to:
reconstruct the block of the current picture of the video data;
determine a classification, from among a plurality of classifications, of the block according to a rule, wherein the rule is adaptively determined based on characteristics of coded sequences of the video data, and wherein the rule is based on a bit-depth of a pixel in the block; and
apply an adaptive loop filter (ALF) to the block based on the classification of the block.
30. An apparatus configured to decode video data, the apparatus comprising:
means for reconstructing a block of a current picture of the video data;
means for determining a classification, from among a plurality of classifications, of the block according to a rule, wherein the rule is adaptively determined based on characteristics of coded sequences of the video data, and wherein the rule is based on a bit-depth of a pixel in the block; and
means for applying an adaptive loop filter (ALF) to the block based on the classification of the block.