LUMA-BASED CHROMA INTRA PREDICTION

• US 20130136174A1
• Filed: 07/12/2011
• Published: 05/30/2013
• Est. Priority Date: 07/12/2011
• Status: Active Grant

• Current Assignee: Intel Corporation
• Original Assignee: Intel Corporation
• Inventors: Xu, Lidong, Chiu, Yi-Jen, Zhang, Wenhao, Han, Yu
• Granted Patent: US 8,724,711 B2
• Time in Patent Office: Days
• US Class Current: 375/240.12
• CPC Class Codes: H04N 19/105 Selection of the reference unit for prediction within a chosen coding or prediction mode, e.g. adaptive choice of position and number of pixels used for prediction
• CPC Class Codes: H04N 19/11 among a plurality of spatial predictive coding modes
• CPC Class Codes: H04N 19/186 the unit being a colour or a chrominance component
• CPC Class Codes: H04N 19/61 in combination with predictive coding
• CPC Class Codes: H04N 19/82 involving filtering within a prediction loop

Abstract

Systems, devices and methods for performing luma-based chroma ultra prediction are described. Down-sample filters may be applied to luma values of luma pixel positions to generate reconstructed luma values for chroma pixel positions in a prediction unit of an intra frame. The down-sampled reconstructed luma values may then be used to predict chroma values for the chroma pixel positions. In some implementations, a reconstructed luma value of a chroma position may be used to predict that position'"'"'s chroma value. In other implementations, reconstructed luma values of neighboring chroma pixel positions may be analyzed to adaptively predict a chroma value for a chroma pixel position.

Claims

1. A computer-implemented method, comprising:

at a video decoder,

determining reconstructed luma values for a first plurality of chroma pixel positions of a current prediction unit (PU) in response to luma values of a plurality of luma pixel positions of the current PU; and

predicting a chroma value of a first chroma pixel position of the first plurality of chroma pixel positions in response to at least one of a reconstructed luma value of the first chroma pixel position or luma values of a second plurality of chroma pixel positions located adjacent to the first chroma pixel position, wherein the second plurality of chroma pixel positions includes at least some of the first plurality of chroma pixel positions.

2. The method of claim 1, wherein predicting the chroma value of the first chroma pixel position in response to luma values of the second plurality of chroma pixel positions comprises:

identifying a second chroma position having a reconstructed hum value closest in value to the reconstructed luma value of the first chroma pixel position by comparing the reconstructed luma value of the first chroma pixel position to the reconstructed luma values of the second plurality of chroma pixel positions, wherein the second plurality of chroma pixel positions includes the second chroma pixel position.

3. The method of claim 2, wherein the chroma value of the first chroma pixel comprises a chroma value of the second chroma pixel position.

4. The method of claim 1, wherein predicting the chroma value of the first chroma pixel position in response to the reconstructed luma value of the first chroma pixel position comprises applying at least one of a linear transform or a non-linear transform to the reconstructed luma value of the first chroma pixel position.

5. The method of claim 4, further comprising:

determining one or more parameters of the linear transform or the non-linear transform in response to at least one of reconstructed chroma values of the second plurality of chroma pixel positions or reconstructed luma values of the second plurality of chroma pixel positions.

6. The method of claim 5, wherein determining the one or more parameters of the linear transform or the non-linear transform comprises determining at least one of a linear least squares value, a non-linear least squares value, or a weighted least squares value of the reconstructed chroma values and/or reconstructed luma values of the second plurality of chroma pixel positions.

7. The method of claim 1, wherein determining reconstructed luma values for the plurality of chroma pixel positions of the current PU in the response to luma values of the plurality of luma pixel positions of the current PU comprises applying a down-sample filter to the luma values of the plurality of luma pixel positions.

8. The method of claim 1, wherein the plurality of luma pixel positions comprise luma pixel positions in at least one of the current PU or a previously predicted PU adjacent to the current PU.

9. The method of claim 1, wherein the current PU comprises a portion of an intra coded image having one of a YUV 4:

2;

0 format or YUV 4;

2;

2 format.

10. A computer-implemented method, comprising:

at a video encoder,

determining, for a chroma pixel position in a current prediction unit (PU) of an image frame, a textural characteristic of a plurality of luma pixel positions physically adjacent to the chroma pixel position in the image frame;

specifying at least one down-sample filter in response to the textural characteristic; and

providing information specifying the down-sample filter.

11. The method of claim 10, further comprising:

determining filter coefficients associated with the down-sample filter; and

providing the filter coefficients with the information specifying the down-sample filter.

12. The method of claim 10, wherein specifying at least one down-sample filter comprises selecting the down-sample filter from a predetermined plurality of down-sample filters.

13. The method of claim 10, wherein determining the textural characteristic comprises performing edge detection based on luma values of the plurality of luma pixel positions.

14. The method of claim 13, wherein the down-sample filter comprises a 1×

2 down-sample filter or a 2×

1 down-sample filter if an edge is detected adjacent to the chroma pixel position, and wherein the down-sample filter comprises a 3×

2 down-sample filter or a 3×

3 down-sample filter if an edge is not detected adjacent to the chroma pixel position.

15. The method of claim 10, wherein the image frame comprises an intra coded frame having one of a YUV 4:

2;

0 format or YUV 4;

2;

2 format.

16. The method of claim 10, wherein providing information specifying the down-sample filter comprises specifying the down-sample filter in at least one header of a bitstream.

17. The method of claim 10, wherein the plurality of luma pixel positions physically adjacent to the chroma pixel position comprise at least one of luma pixel positions of the current PU or of a PU adjacent to the current PU.

18. A video decoder, comprising:

an intra prediction module configured to;

determine reconstructed luma values for a first plurality of chroma pixel positions of a current prediction unit (PU) in response to luma values of a plurality of luma pixel positions of the current PU; and

predict a chroma value of a first chroma pixel position of the first plurality of chroma pixel positions in response to at least one of a reconstructed luma value of the first chroma pixel position or luma values of a second plurality of chroma pixel positions located adjacent to the first chroma pixel position, wherein the second plurality of chroma pixel positions includes at least some of the first plurality of chroma pixel positions.

19. The video decoder of claim 18, wherein the intra prediction module is configured to predict the chroma value of the first chroma pixel position in response to luma values of the second plurality of chroma pixel positions by identifying a second chroma position having a reconstructed luma value closest in value to the reconstructed luma value of the first chroma pixel position by comparing the reconstructed luma value of the first chroma pixel position to the reconstructed luma values of the second plurality of chroma pixel positions, wherein the second plurality of chroma pixel positions includes the second chroma pixel position.

20. The video decoder of claim 19, wherein the chroma value of the first chroma pixel comprises a chroma value of the second chroma pixel position.

21. The video decoder of claim 18, wherein the intra prediction module is configured to predict the chroma value of the first chroma pixel position in response to the reconstructed luma value of the first chroma pixel position by applying at least one of a linear transform or a non-linear transform to the reconstructed luma value of the first chroma pixel position.

22. The video decoder of claim 21 wherein the intra prediction module is further configured to:

determine one or more parameters of the linear transform or the non-linear transform in response to at least one of reconstructed chroma values of the second plurality of chroma pixel positions or reconstructed luma values of the second plurality of chroma pixel positions.

23. The video decoder of claim 22, wherein the intra prediction module is configured to determine the one or more parameters of the linear transform or the non-linear transform by determining at least one of a linear least squares value, a non-linear least squares value, or a weighted least squares value of the reconstructed chroma values and/or reconstructed luma values of the second plurality of chroma pixel positions.

24. The video decoder of claim 18, wherein the intra prediction module is configured to determine reconstructed luma values for the plurality of chroma pixel positions of the current PU in the response to luma values of the plurality of luma pixel positions of the current PU by applying a down-sample filter to the luma values of the plurality of luma pixel positions.

25. A video encoder, comprising:

an intra prediction module configured to;

determine, for a chroma pixel position in a current prediction unit (PU) of an image frame, a textural characteristic of a plurality of luma pixel positions physically adjacent to the chroma pixel position in the image frame;

specify at least one down-sample filter in response to the textural characteristic; and

provide information specifying the down-sample filter.

26. The video encoder of claim 25, wherein the intra prediction module is further configured to:

determine filter coefficients associated with the down-sample filter; and

provide the filter coefficients with the information specifying the down-sample filter.

27. The video encoder of claim 25, wherein the intra prediction module is configured to determine the textural characteristic by performing edge detection based on luma values of the plurality of luma pixel positions.

28. The video encoder of claim 27, wherein the down-sample filter comprises a 1×

2 down-sample filter or a 2×

1 down-sample filter if an edge is detected adjacent to the chroma pixel position, and wherein the down-sample filter comprises a 3×

2 down-sample filter or a 3×

3 down-sample filter if an edge is not detected adjacent to the chroma pixel position.

29. The video encoder of claim 25, wherein the intra prediction module is configured to provide information specifying the down-sample filter by specifying the down-sample filter in at least one header of a bitstream.

30. An article comprising a computer program product having stored therein instructions that, if executed, result in:

at a video decoder,

determining reconstructed luma values for a first plurality of chroma pixel positions of a current prediction unit (PU) in response to luma values of a plurality of luma pixel positions of the current PU; and

predicting a chroma value of a first chroma pixel position of the first plurality of chroma pixel positions in response to at least one of a reconstructed luma value of the first chroma pixel position or luma values of a second plurality of chroma pixel positions located adjacent to the first chroma pixel position, wherein the second plurality of chroma pixel positions includes at least some of the first plurality of chroma pixel positions.

31. The article of claim 30, wherein predicting the chroma value of the first chroma pixel position in response to luma values of the second plurality of chroma pixel positions comprises:

identifying a second chroma position having a reconstructed luma value closest in value to the reconstructed luma value of the first chroma pixel position by comparing the reconstructed luma value of the first chroma pixel position to the reconstructed luma values of the second plurality of chroma pixel positions, wherein the second plurality of chroma pixel positions includes the second chroma pixel position.

32. The article of claim 31, wherein the chroma value of the first chroma pixel comprises a chroma value of the second chroma pixel position.

33. The article of claim 30, wherein predicting the chroma value of the first chroma pixel position in response to the reconstructed luma value of the first chroma pixel position comprises applying at least one of a linear transform or a non-linear transform to the reconstructed luma value of the first chroma pixel position.

34. The article of claim 33, further comprising instructions that, if executed, result in:

determining one or more parameters of the linear transform or the non-linear transform in response to at least one of reconstructed chroma values of the second plurality of chroma pixel positions or reconstructed luma values of the second plurality of chroma pixel positions.

35. The article of claim 34, wherein determining the one or more parameters of the linear transform or the non-linear transform comprises determining at least one of a linear least squares value, a non-linear least squares value, or a weighted least squares value of the reconstructed chroma values and/or reconstructed luma values of the second plurality of chroma pixel positions.

36. The article of claim 30, wherein determining reconstructed luma values for the plurality of chroma pixel positions of the current PU in the response to luma values of the plurality of luma pixel positions of the current PU comprises applying a down-sample filter to the luma values of the plurality of luma pixel positions.