Enhanced multiple transforms for prediction residual

• US 10,306,229 B2
• Filed: 01/25/2016
• Issued: 05/28/2019
• Est. Priority Date: 01/26/2015
• Status: Active Grant

• Current Assignee: Qualcomm, Inc.
• Original Assignee: Qualcomm, Inc.
• Inventors: Zhao, Xin, Lee, Sungwon, Chen, Jianle, Zhang, Li, Li, Xiang, Chen, Ying, Karczewicz, Marta, Liu, Hongbin
• Primary Examiner(s): Werner, David N
• Application Number:
  US15/005,736
• Publication Number:
  US 20160219290A1
• Time in Patent Office: 1,219 Days
• CPC Class Codes: H04N 19/103 Selection of coding mode or of prediction mode
• CPC Class Codes: H04N 19/109 among a plurality of temporal predictive coding modes
• CPC Class Codes: H04N 19/112 according to a given display mode, e.g. for interlaced or progressive display mode
• CPC Class Codes: H04N 19/12 Selection from among a plurality of transforms or standards, e.g. selection between discrete cosine transform [DCT] and sub-band transform or selection between H.263 and H.264
• CPC Class Codes: H04N 19/122 Selection of transform size, e.g. 8x8 or 2x4x8 DCT; Selection of sub-band transforms of varying structure or type
• CPC Class Codes: H04N 19/129 Scanning of coding units, e.g. zig-zag scan of transform coefficients or flexible macroblock ordering [FMO]
• CPC Class Codes: H04N 19/136 Incoming video signal characteristics or properties
• CPC Class Codes: H04N 19/157 Assigned coding mode, i.e. the coding mode being predefined or preselected to be further used for selection of another element or parameter
• CPC Class Codes: H04N 19/159 Prediction type, e.g. intra-frame, inter-frame or bidirectional frame prediction
• CPC Class Codes: H04N 19/167 Position within a video image, e.g. region of interest [ROI]
• CPC Class Codes: H04N 19/176 the region being a block, e.g. a macroblock
• CPC Class Codes: H04N 19/60 using transform coding
• CPC Class Codes: H04N 19/70 characterised by syntax aspects related to video coding, e.g. related to compression standards

Abstract

Example techniques are described to determine transforms to be used during video encoding and video decoding. A video encoder and a video decoder may select transform subsets that each identify one or more candidate transforms. The video encoder and the video decoder may determine transforms from the selected transform subsets.

Claims

1. A method of decoding video data, the method comprising:

determining a plurality of transform subsets for transforming a current coefficient block of a video block encoded according to one of a plurality of prediction modes, the prediction modes comprising a plurality of intra-prediction modes and a plurality of inter prediction modes, each subset identifying one or more candidate transforms, wherein at least one transform subset identifies a plurality of candidate transforms, wherein determining the plurality of transform subsets comprises determining either a first, same, one of the plurality of transform subsets for each of the intra-prediction modes, or a second, same, one of the plurality of transform subsets for each of the inter-prediction modes, wherein the plurality of transform subsets is the same for each of the intra-prediction modes, and wherein the plurality of transform subsets is the same for each of the inter-prediction modes;

selecting a first transform subset from the determined plurality of transform subsets for a first transform for the current coefficient block of the video data;

selecting a second transform subset from the determined plurality of transform subsets for a second transform for the current coefficient block of the video data;

determining the first transform from the selected first transform subset;

determining the second transform from the selected second transform subset;

determining a current transform block based on applying, to coefficients in the current coefficient block, a basis function defined by the first transform and a basis function defined by the second transform; and

reconstructing the video block based on an addition of values of the current transform block to corresponding values of a predictive block.

2. The method of claim 1, further comprising:

determining an intra-prediction mode of the video block based on a prediction mode of the video block being intra-prediction,

wherein selecting the first transform subset comprises selecting the first transform subset based on the determined intra-prediction mode,

wherein selecting the second transform subset comprises selecting the second transform subset based on the determined intra-prediction mode, and

wherein reconstructing the video block comprises reconstructing the video block based on the determined intra-prediction mode.

3. The method of claim 1, further comprising:

determining a location of the current transform block in the video block based on a prediction mode of the video block being inter-prediction,

wherein selecting the first transform subset comprises selecting the first transform subset based on the determined location of the current transform block, and

wherein selecting the second transform subset comprises selecting the second transform subset based on the determined location of the current transform block.

4. The method of claim 1, further comprising:

receiving a first transform subset index into the first transform subset; and

receiving a second transform subset index into the second transform subset,

wherein determining the first transform comprises determining the first transform based on a transform in the first transform subset identified by the first transform subset index, and

wherein determining the second transform comprises determining the second transform based on a transform in the second transform subset identified by the second transform subset index.

5. The method of claim 1, further comprising:

determining that a number of nonzero coefficients in the current coefficient block is less than a threshold,

wherein determining the first transform comprises determining that a particular transform identified in the first transform subset is the first transform without receiving a transform subset index into the first transform subset, in response to determining that the number of nonzero coefficients in the current coefficient block is less than the threshold, and

wherein determining the second transform comprises determining that a particular transform identified in the second transform subset is the second transform without receiving a transform subset index into the second transform subset, in response to determining that the number of nonzero coefficients in the current coefficient block is less than the threshold.

6. The method of claim 1, further comprising:

receiving a flag indicating that not all transform blocks of the video block that includes the current transform block are transformed using the same transform,

wherein selecting the first transform subset for the first transform for the current coefficient block of the video data, selecting the second transform subset for the second transform for the current coefficient block of the video data, determining the first transform from the selected first transform subset, and determining the second transform from the selected second transform subset comprise selecting the first transform subset for the first transform for the current coefficient block of the video data, selecting the second transform subset for the second transform for the current coefficient block of the video data, determining the first transform from the selected first transform subset, and determining the second transform from the selected second transform subset, in response to receiving the flag indicating that not all transform blocks of the video block that includes the current transform block are transformed using the same transform.

7. The method of claim 6, wherein the video block comprises one of:

a coding tree unit (CTU), a coding unit (CU), or a prediction unit (PU).

8. The method of claim 1, wherein at least one of the first transform subset or the second transform subset includes a transform that is different than a discrete cosine transform (DCT)-II transform and a discrete sine transform (DST)-VII transform.

9. The method of claim 1, wherein the first transform subset and the second transform subset include different transform types.

10. The method of claim 1, wherein the plurality of transform subsets comprises three or more transform subsets.

11. The method of claim 1, wherein the candidate transforms are different transform types.

12. The method of claim 1, wherein determining the plurality of transform subsets comprises determining the plurality of transform subsets based on a size of the video block.

13. The method of claim 1, further comprising:

receiving from a bitstream information indicating a prediction mode;

receiving from the bitstream information indicating coefficients of the current coefficient block;

constructing the current coefficient block based on the received information indicating the coefficients; and

determining the predictive block based on the prediction mode.

14. The method of claim 13, wherein the prediction mode comprises an inter-prediction mode.

15. The method of claim 1, wherein the current transform block is a residual of the video block and the predictive block.

16. A method of encoding video data, the method comprising:

determining a plurality of transform subsets for transforming a current transform block of a video block encoded according to one of a plurality of prediction mode, the prediction modes comprising a plurality of intra-prediction modes and a plurality of inter-prediction modes, each subset identifying one or more candidate transforms, wherein at least one transform subset identifies a plurality of candidate transforms, wherein determining the plurality of transform subsets comprises determining either a first, same, one of the plurality transform subsets for each of the intra-prediction modes, or a second, same, one of the plurality of transform subsets for each of the inter-prediction modes, wherein the plurality of transform subsets is the same for each of the intra-prediction modes, and wherein the plurality of transform subsets is the same for each of the inter-prediction modes;

selecting a first transform subset from the determined plurality of transform subsets for a first transform for the current transform block of the video block of the video data;

selecting a second transform subset from the determined plurality of transform subsets for a second transform for the current transform block of the video block of the video data;

determining the first transform from the selected first transform subset;

determining the second transform from the selected second transform subset;

determining a current coefficient block based on applying, to the current transform block, a basis function defined by the first transform and a basis function defined by the second transform; and

generating a video bitstream that includes information indicative of coefficients of the current coefficient block used for reconstruction of the video block.

17. The method of claim 16, further comprising:

determining an intra-prediction mode of the video block,

wherein selecting the first transform subset comprises selecting the first transform subset based on the determined intra-prediction mode, and

wherein selecting the second transform subset comprises selecting the second transform subset based on the determined intra-prediction mode.

18. The method of claim 16, further comprising:

determining a location of the current transform block in the video block based on the video block being inter-prediction encoded,

wherein selecting the first transform subset comprises selecting the first transform subset based on the determined location of the current transform block, and

wherein selecting the second transform subset comprises selecting the second transform subset based on the determined location of the current transform block.

19. The method of claim 16, further comprising:

generating in the video bitstream a first transform subset index into the first transform subset to identify a transform in the first transform subset used to determine the current coefficient block; and

generating in the video bitstream a second transform subset index into the second transform subset to identify a transform in the second transform subset used to determine the current coefficient block.

20. The method of claim 19, further comprising:

determining a number of nonzero coefficients in the current coefficient block,

wherein signaling the first transform subset index comprises signaling the first transform subset index based on the number of nonzero coefficients being greater than a threshold, and

wherein signaling the second transform subset index comprises signaling the second transform subset index based on the number of nonzero coefficients being greater than the threshold.

21. The method of claim 16, wherein at least one of the first transform subset or the second transform subset includes a transform that is different than a discrete cosine transform (DCT)-II transform and a discrete sine transform (DST)-VII transform.

22. The method of claim 16, wherein the first transform subset and the second transform subset include different transform types.

23. The method of claim 16, wherein the plurality of transform subsets comprises three or more transform subsets.

24. The method of claim 16, wherein the candidate transforms are different transform types.

25. The method of claim 16, wherein determining the plurality of transform subsets comprises determining the plurality of transform subsets based on a size of the video block.

26. The method of claim 16, further comprising:

determining the predictive block; and

generating in the video bitstream information indicative of a prediction mode of the video block based on the predictive block.

27. The method of claim 26, wherein the predictive block is a block in the same picture as the video block based on the video block being intra-predicted or in a picture different than the picture that includes the video block based on the video block being inter-predicted.

28. The method of claim 16, further comprising:

determining the current transform block as a residual between the video block and the predictive block.

29. A device for video decoding video data, the device comprising:

a video data memory configured to store the video data and transform subsets, each subset identifying one or more candidate transforms, wherein at least one transform subset identifies a plurality of candidate transforms; and

a video decoder comprising integrated circuitry, the video decoder configured to;

determine a plurality of transform subsets from the stored transform subsets for transforming a current coefficient block of a video block encoded according to one of a plurality of prediction modes, the prediction modes comprising a plurality of intra-prediction modes and a plurality of inter-prediction modes, wherein to determine the plurality of transform subsets, the video decoder is configured to determine either a first, same, one of the plurality of transform subsets for each of the intra-prediction modes or a second, same, one of the plurality of transform subsets for each of the inter-prediction modes, wherein the plurality of transform subsets is the same for each of the intra-prediction modes, and wherein the plurality of transform subsets is the same for each of the inter-prediction modes;

select a first transform subset from the determined plurality of transform subsets for a first transform for the current coefficient block of the video data;

select a second transform subset from the determined plurality of transform subsets for a second transform for the current coefficient block of the video data;

determine the first transform from the selected first transform subset;

determine the second transform from the selected second transform subset;

determine a current transform block based on applying, to the coefficients in the current coefficient block, a basis function defined by the first transform and a basis function defined by the second transform; and

reconstruct the video block based on an addition of values of the current transform block to corresponding values of a predictive block.

30. The device of claim 29,

wherein the video decoder is configured to determine an intra-prediction mode of the video block based on a prediction mode of the video block being intra-prediction,

wherein to select the first transform subset, the video decoder is configured to select the first transform subset based on the determined intra-prediction mode,

wherein to select the second transform subset, the video decoder is configured to select the second transform subset based on the determined intra-prediction mode, and

wherein to reconstruct the video block, the video decoder is configured to reconstruct the video block based on the determined intra-prediction mode.

31. The device of claim 29,

wherein the video decoder is configured to determine a location of the current transform block in the video block based on a prediction mode of the video block being inter-prediction,

wherein to select the first transform subset, the video decoder is configured to select the first transform subset based on the determined location of the current transform block, and

wherein to select the second transform subset, the video decoder is configured to select the second transform subset based on the determined location of the current transform block.

32. The device of claim 29, wherein the video decoder is configured to:

receive a first transform subset index into the first transform subset; and

receive a second transform subset index into the second transform subset,

wherein to determine the first transform, the video decoder is configured to determine the first transform based on a transform in the first transform subset identified by the first transform subset index, and

wherein to determine the second transform, the video decoder is configured to determine the second transform based on a transform in the second transform subset identified by the second transform subset index.

33. The device of claim 29, wherein the video decoder is configured to:

determine that a number of nonzero coefficients in the current coefficient block is less than a threshold,

wherein to determine the first transform, the video decoder is configured to determine that a particular transform identified in the first transform subset is the first transform without receiving a transform subset index into the first transform subset, in response to determining that the number of nonzero coefficients in the current coefficient block is less than the threshold, and

wherein to determine the second transform, the video decoder is configured to determine that a particular transform identified in the second transform subset is the second transform without receiving a transform subset index into the second transform subset, in response to determining that the number of nonzero coefficients in the current coefficient block is less than the threshold.

34. The device of claim 29, wherein the video decoder is configured to:

receive a flag indicating that not all transform blocks of the video block that includes the current transform block are transformed using the same transform,

wherein to select the first transform subset for the first transform for the current coefficient block of the video data, select the second transform subset for the second transform for the current coefficient block of the video data, determine the first transform from the selected first transform subset, and determine the second transform from the selected second transform subset, the video decoder is configured to select the first transform subset for the first transform for the current coefficient block of the video data, select the second transform subset for the second transform for the current coefficient block of the video data, determine the first transform from the selected first transform subset, and determine the second transform from the selected second transform subset, in response to receiving the flag indicating that not all transform blocks of the video block that includes the current transform block are transformed using the same transform.

35. The device of claim 34, wherein the video block comprises one of:

a coding tree unit (CTU), a coding unit (CU), or a prediction unit (PU).

36. The device of claim 29, wherein at least one of the first transform subset or the second transform subset includes a transform that is different than a discrete cosine transform (DCT)-II transform and a discrete sine transform (DST)-VII transform.

37. The device of claim 29, wherein the first transform subset and the second transform subset include different transform types.

38. The device of claim 29, wherein the plurality of transform subsets comprises three or more transform subsets.

39. The device of claim 29, wherein the candidate transforms are different transform types.

40. The device of claim 29, wherein to determine the plurality of transform subsets, the video decoder is configured to determine the plurality of transform subsets based on a size of the video block.

41. The device of claim 29, wherein the video decoder is configured to:

receive from a bitstream information indicating a prediction mode;

receive from the bitstream information indicating coefficients of the current coefficient block;

construct the current coefficient block based on the received information indicating the coefficients; and

determine the predictive block based on the prediction mode.

42. The device of claim 41, wherein the prediction mode comprises an intra-prediction mode.

43. The device of claim 29, wherein the current transform block is a residual of the video block and the predictive block.

44. A device for encoding video data, the device comprising:

a video data memory configured to store the video data and transform subsets, each subset identifying one or more candidate transforms, wherein at least one transform subset identifies a plurality of candidate transforms; and

a video encoder configured to;

determine a plurality of transform subsets from the stored transform subsets for transforming a current transform block of a video block encoded according to one of a plurality of prediction mode, the prediction modes comprising a plurality of intra-prediction modes and a plurality of inter-prediction modes, wherein to determine the plurality of transform subsets the video encoder is configured to determine either a first, same, one of the plurality transform subsets for each of the intra-prediction modes, or a second, same, one of the plurality of transform subsets for each of the inter-prediction modes, wherein the plurality of transform subsets is the same for each of the intra-prediction modes, and wherein the plurality of transform subsets is the same for each of the inter-prediction modes;

select a first transform subset from the determined plurality of transform subsets for a first transform for the current transform block of the video block of the video data;

select a second transform subset from the determined plurality of transform subsets for a second transform for the current transform block of the video block of the video data;

determine the first transform from the selected first transform subset;

determine the second transform from the selected second transform subset;

determine a current coefficient block based on applying, to the current transform block, a basis function defined by the first transform and a basis function defined by the second transform; and

generate a video bitstream that includes information indicative of coefficients of the current coefficient block used for reconstruction of the video block.

45. The device of claim 44,

wherein the video encoder is configured to determine an intra-prediction mode of the video block,

wherein to select the first transform subset, the video encoder is configured to select the first transform subset based on the determined intra-prediction mode, and

wherein to select the second transform subset, the video encoder is configured to select the second transform subset based on the determined intra-prediction mode.

46. The device of claim 44,

wherein the video encoder is configured to determine a location of the current transform block in the video block based on the video block being inter-prediction encoded,

wherein to select the first transform subset, the video encoder is configured to select the first transform subset based on the determined location of the current transform block, and

wherein to select the second transform subset, the video encoder is configured to select the second transform subset based on the determined location of the current transform block.

47. The device of claim 44, wherein the video encoder is configured to:

generate in the video bitstream a first transform subset index into the first transform subset to identify a transform in the first transform subset used to determine the current coefficient block; and

generate in the video bitstream a second transform subset index into the second transform subset to identify a transform in the second transform subset used to determine the current coefficient block.

48. The device of claim 47,

wherein the video encoder is configured to determine a number of nonzero coefficients in the current coefficient block,

wherein to generate in the video bitstream the first transform subset index, the video encoder is configured to generate in the video bitstream the first transform subset index based on the number of nonzero coefficients being greater than a threshold, and

wherein to generate the second transform subset index, the video encoder is configured to generate in the video bitstream the second transform subset index based on the number of nonzero coefficients being greater than the threshold.

49. The device of claim 44, wherein at least one of the first transform subset or the second transform subset includes a transform that is different than a discrete cosine transform (DCT)-II transform and a discrete sine transform (DST)-VII transform.

50. The device of claim 44, wherein the first transform subset and the second transform subset include different transform types.

51. The device of claim 44, wherein the plurality of transform subsets comprises three or more transform subsets.

52. The device of claim 44, wherein the candidate transforms are different transform types.

53. The device of claim 44, wherein to determine the plurality of transform subsets, the video encoder is configured to determine the plurality of transform subsets based on a size of the video block.

54. The device of claim 44, wherein the video encoder is configured to:

determine the predictive block; and

generate in the video bitstream information indicative of a prediction mode of the video block based on the predictive block.

55. The device of claim 44, wherein the predictive block is a block in the same picture as the video block based on the video block being intra-predicted or in a picture different than the picture that includes the video block based on the video block being inter-predicted.

56. The device of claim 44, wherein the video encoder is configured to:

determine the current transform block as a residual between the video block and the predictive block.

57. A device for decoding video data, the device comprising:

means for determining a plurality of transform subsets for transforming a current coefficient block of a video block encoded according to one of a plurality of prediction modes, the prediction modes comprising a plurality of intra-prediction modes and a plurality of inter prediction modes, each subset identifying one or more candidate transforms, wherein at least one transform subset identifies a plurality of candidate transforms, wherein the means for determining the plurality of transform subsets comprises means for determining either a first, same, one of the plurality of transform subsets for each of the intra-prediction modes, or a second, same, one of the plurality of transform subsets for each of the inter-prediction modes, wherein the plurality of transform subsets is the same for each of the intra-prediction modes, and wherein the plurality of transform subsets is the same for each of the inter-prediction modes;

means for selecting a first transform subset from the determined plurality of transform subsets for a first transform for the current coefficient block of the video data;

means for selecting a second transform subset from the determined plurality of transform subsets for a second transform for the current coefficient block of the video data;

means for determining the first transform from the selected first transform subset;

means for determining the second transform from the selected second transform subset;

means for determining a current transform block based on applying, to coefficients in the current coefficient block, a basis function defined by the first transform and a basis function defined by the second transform; and

means for reconstructing the video block based on an addition of values of the current transform block to corresponding values of a predictive block.

58. A non-transitory computer-readable storage medium storing instructions that when executed cause a video decoder of a device for video decoding to:

determine a plurality of transform subsets for transforming a current coefficient block of a video block encoded according to one of a plurality of prediction modes, the prediction modes comprising a plurality of intra-prediction modes and a plurality of inter prediction modes, each subset identifying one or more candidate transforms, wherein at least one transform subset identifies a plurality of candidate transforms, wherein the instructions that cause the video decoder to determine the plurality of transform subsets comprise instructions that cause the video decoder to determine either a first, same, one of the plurality of transform subsets for each of the intra-prediction modes, or a second, same, one of the plurality of transform subsets for each of the inter-prediction modes, wherein the plurality of transform subsets is the same for each of the intra-prediction modes, and wherein the plurality of transform subsets is the same for each of the inter-prediction modes;

select a first transform subset from the determined plurality of transform subsets for a first transform for the current coefficient block of the video data;

select a second transform subset from the determined plurality of transform subsets for a second transform for the current coefficient block of the video data;

determine the first transform from the selected first transform subset;

determine the second transform from the selected second transform subset;

determine a current transform block based on applying, to coefficients in the current coefficient block, a basis function defined by the first transform and a basis function defined by the second transform; and

reconstruct the video block based on an addition of values of the current transform block to corresponding values of a predictive block.

59. A device for encoding video data, the device comprising:

means for determining a plurality of transform subsets for transforming a current transform block of a video block encoded according to one of a plurality of prediction mode, the prediction modes comprising a plurality of intra-prediction modes and a plurality of inter-prediction modes, each subset identifying one or more candidate transforms, wherein at least one transform subset identifies a plurality of candidate transforms, wherein the means for determining the plurality of transform subsets comprises means for determining either a first, same, one of the plurality transform subsets for each of the intra-prediction modes, or a second, same, one of the plurality of transform subsets for each of the inter-prediction modes, wherein the plurality of transform subsets is the same for each of the intra-prediction modes, and wherein the plurality of transform subsets is the same for each of the inter-prediction modes;

means for selecting a first transform subset from the determined plurality of transform subsets for a first transform for the current transform block of the video block of the video data;

means for selecting a second transform subset from the determined plurality of transform subsets for a second transform for the current transform block of the video block of the video data;

means for determining the first transform from the selected first transform subset;

means for determining the second transform from the selected second transform subset;

means for determining a current coefficient block based on applying, to the current transform block, a basis function defined by the first transform and a basis function defined by the second transform; and

means for generating a video bitstream that includes information indicative of coefficients of the current coefficient block used for reconstruction of the video block.

60. A non-transitory computer-readable storage medium storing instructions that when executed cause a video encoder of a device for video encoding to:

determine a plurality of transform subsets for transforming a current transform block of a video block encoded according to one of a plurality of prediction mode, the prediction modes comprising a plurality of intra-prediction modes and a plurality of inter-prediction modes, each subset identifying one or more candidate transforms, wherein at least one transform subset identifies a plurality of candidate transforms, wherein the instructions that cause the video encoder to determine the plurality of transform subsets comprise instructions that cause the video encoder to determine either a first, same, one of the plurality transform subsets for each of the intra-prediction modes, or a second, same, one of the plurality of transform subsets for each of the inter-prediction modes, wherein the plurality of transform subsets is the same for each of the intra-prediction modes, and wherein the plurality of transform subsets is the same for each of the inter-prediction modes;

select a first transform subset from the determined plurality of transform subsets for a first transform for the current transform block of the video block of the video data;

select a second transform subset from the determined plurality of transform subsets for a second transform for the current transform block of the video block of the video data;

determine the first transform from the selected first transform subset;

determine the second transform from the selected second transform subset;

determine a current coefficient block based on applying, to the current transform block, a basis function defined by the first transform and a basis function defined by the second transform; and

generate a video bitstream that includes information indicative of coefficients of the current coefficient block used for reconstruction of the video block.