Non-square transform units and prediction units in video coding

Patent metadata

Field	Value
Publication number	US-9787982-B2
Application number	US-201213608380-A
Country	US
Kind code	B2
Filing date	Sep 10, 2012
Priority date	Sep 12, 2011
Publication date	Oct 10, 2017
Grant date	Oct 10, 2017

Abstract

This disclosure proposes techniques for transform partitioning in an intra-prediction video coding process. In one example, for a given intra-predicted block, a reduced number of transform unit partition options is allowed, based on certain conditions. In another example, transform units are decoupled from prediction units for intra-predicted block. For a given prediction unit, transforms of different sizes and shapes from the prediction unit may be applied. In another example, a reduced number of intra-prediction modes are allowed for a prediction unit having a non-square shape.

First claim

What is claimed is:

1. A video encoding method comprising:
determining an intra-prediction mode from a set of intra-prediction modes for a prediction unit of video data, wherein the set of intra-prediction modes includes at least a first intra-prediction mode and a second intra-prediction mode, the second intra-prediction mode is different than the first intra-prediction mode, and wherein each of the first intra-prediction mode and the second intra-prediction mode is at least one of a directional intra-prediction mode, a direct current (DC) intra-prediction mode, or a planar mode;
determining a subset of transform unit partitions from a set of transform unit partitions based on the determined intra-prediction mode, such that a first subset of transform partitions is determined if the determined intra-prediction mode is the first intra-prediction mode, and a second subset of transform partitions is determined if the determined intra-prediction mode is the second intra-prediction mode, wherein the set of transform unit partitions includes at least the first subset of transform partitions and the second subset of transform partitions, and wherein the second subset of transform partitions is different than the first subset of transform partitions, wherein when:
the determined intra-prediction mode is a vertically-oriented intra-prediction mode, the subset of transform unit partitions includes vertically-oriented non-square transform unit partitions and does not include at least one horizontally-oriented non-square transform unit partition,
the determined intra-prediction mode is a horizontally-oriented intra-prediction mode, the subset of transform unit partitions includes horizontally-oriented non-square transform unit partitions and does not include at least one vertically-oriented non-square transform unit partition,
the determined intra-prediction mode is a DC mode, a planar mode, or a diagonal mode, the subset of transform unit partitions includes square transform unit partitions and does not include at least one non-square transform unit partition;

determining a transform unit partition for the prediction unit of video data from the subset of transform unit partitions; and
encoding the prediction unit of video data using the determined transform unit partition.

2. The video encoding method of claim 1 , wherein the subset of transform unit partitions only includes one transform unit partition per intra-prediction mode.

3. The video encoding method of claim 1 , wherein an equal number of transform unit partitions are present in each subset of transform unit partitions for each intra-prediction mode in the set of intra-prediction modes.

4. The video encoding method of claim 1 ,
wherein a partition of the prediction has a different size and shape than the determined transform unit partition.

5. The video encoding method of claim 1 , further comprising:
performing intra-prediction according to the determined intra-prediction mode on the prediction unit of video data to produce a prediction residual,
wherein determining the transform unit partition comprises:
testing a plurality of transform partitions of the determined subset of transform unit partitions on the prediction residual; and
selecting one of the plurality of transform partitions as the determined transform partition based on the testing.

6. The video encoding method of claim 1 , further comprising:
signaling the determined transform partition in an encoded video bitstream.

7. The video encoding method of claim 1 , further comprising:
determining a prediction unit partition from among a plurality of prediction unit partitions for the prediction unit of video data; and
determining the set of intra-prediction modes based on the determined prediction unit partition.

8. The video encoding method of claim 7 , wherein the determined prediction unit partition is a vertically-oriented non-square prediction unit partition, and wherein the set of intra-prediction modes includes vertically-oriented intra-prediction modes and does not include at least one horizontally-oriented non-square transform unit partition.

9. The video encoding method of claim 7 , wherein the determined prediction unit partition is a horizontally-oriented non-square prediction unit partition, and wherein the set of intra-prediction modes includes horizontally-oriented intra-prediction modes and does not include at least one vertically-oriented non-square transform unit partition.

10. A video decoding method comprising:
receiving an indication of an intra-prediction mode from among a set of intra-prediction modes for a prediction unit of video data, wherein the set of intra-prediction modes includes at least a first intra-prediction mode and a second intra-prediction mode, the second intra-prediction mode is different than the first intra-prediction mode, and wherein each of the first intra-prediction mode and the second intra-prediction mode is at least one of a directional intra-prediction mode, a direct current (DC) intra-prediction mode, or a planar mode;
determining a subset of transform unit partitions from a set of transform unit partitions based on the determined intra-prediction mode, such that a first subset of transform partitions is determined if the determined intra-prediction mode is the first intra-prediction mode, and a second subset of transform partitions is determined if the determined intra-prediction mode is the second intra-prediction mode, wherein the set of transform unit partitions includes at least the first subset of transform partitions and the second subset of transform partitions, and wherein the second subset of transform partitions is different than the first subset of transform partitions, wherein when:
the determined intra-prediction mode is a vertically-oriented intra-prediction mode, the subset of transform unit partitions includes vertically-oriented non-square transform unit partitions and does not include at least one horizontally-oriented non-square transform unit partition,
the determined intra-prediction mode is a horizontally-oriented intra-prediction mode, the subset of transform unit partitions includes horizontally-oriented non-square transform unit partitions and does not include at least one vertically-oriented non-square transform unit partition,
the determined intra-prediction mode is a DC mode, a planar mode, or a diagonal mode, the subset of transform unit partitions includes square transform unit partitions and does not include at least one non-square transform unit partition;

determining a transform unit partition for the prediction unit of video data from the subset of transform unit partitions; and
decoding the prediction unit of video data using the determined transform unit partition.

11. The video decoding method of claim 10 , wherein the subset of transform unit partitions only includes one transform unit partition per intra-prediction mode.

12. The video decoding method of claim 10 , wherein an equal number of transform unit partitions are present in each subset of transform unit partitions for each intra-prediction mode in the set of intra-prediction modes.

13. The video decoding method of claim 10 , wherein a partition of the prediction unit has a different size and shape than the determined transform unit partition.

14. The video decoding method of claim 10 , further comprising:
receiving an indication of a prediction unit partition from among a plurality of prediction unit partitions for the prediction unit of video data; and
determining the set of intra-prediction modes based on the prediction unit partition.

Assignees

• Wang Xianglin
• Karczewicz Marta
• Guo Liwei
• Qualcomm Inc

Inventors

• Wang Xianglin
• Karczewicz Marta
• Guo Liwei

Classifications

• H04N19/157

  Assigned coding mode, i.e. the coding mode being predefined or preselected to be further used for selection of another element or parameter · CPC title

• H04N19/61 Primary

  in combination with predictive coding · CPC title

• H04N19/122 Primary

  Selection of transform size, e.g. 8x8 or 2x4x8 DCT; Selection of sub-band transforms of varying structure or type · CPC title

• H04N19/176

  the region being a block, e.g. a macroblock · CPC title

• H04N19/11

  among a plurality of spatial predictive coding modes · CPC title