Cross-channel residual prediction

Patent metadata

Field	Value
Publication number	US-10015485-B2
Application number	US-201113977420-A
Country	US
Kind code	B2
Filing date	Nov 7, 2011
Priority date	Nov 7, 2011
Publication date	Jul 3, 2018
Grant date	Jul 3, 2018

Abstract

Systems, apparatus and methods are described including determining a prediction residual for a channel of video data; and determining, using the first channel's prediction residual, a prediction residual for a second channel of the video data. Further, a prediction residual for a third channel of the video data may be determined using the second channel's prediction residual.

First claim

What is claimed:

1. A computer-implemented method, comprising:
at a video encoder:
dividing a codec dependent range of possible reconstructed prediction residual values for a first channel of video data into a plurality of subsets of possible reconstructed prediction residual values;
assigning each of the subsets a first linear model parameter and a second linear model parameter for linear cross-channel residual prediction;
determining a first reconstructed prediction residual value at a particular pixel position within a picture having a plurality of pixel positions for the first channel of video data, wherein the particular position is within a first block of the video data;
determining, using the first reconstructed prediction residual value, a first predicted residual value at the particular pixel position for a second channel of the video data, wherein determining the first predicted residual value comprises:
determining the first reconstructed prediction residual value at the particular pixel position for the first channel is within a particular subset of the subsets of possible reconstructed prediction residual values; and
predicting the first predicted residual value for the second channel based on first and second linear model parameters corresponding to the particular subset of the subsets of possible reconstructed prediction residual values;

determining second reconstructed prediction residual values for a second block of the first channel of the video data and third reconstructed prediction residual values for a third block of the first channel of the video data, wherein the second block is intra coded and the third block is inter coded;
determining, based on the second reconstructed prediction residual values, second predicted residual values for the second channel of the video data using linear cross-channel residual prediction, wherein the linear cross-channel residual prediction is performed responsive to the second block being intra coded; and
skipping cross-channel residual prediction for the second channel of the video data for the third block responsive to the third block being inter coded.

2. The method of claim 1 , wherein when the first channel comprises a luma channel, the second channel comprises a chroma channel and when the first channel comprises a chroma channel, the second channel comprises one of a luma channel or another chroma channel.

3. The method of claim 1 , further comprising:
transforming and quantizing the first predicted residual value for the second channel and inverse quantizing and inverse transforming the transformed and quantized first predicted residual value for the second channel to generate a second reconstructed prediction residual value for the second channel; and
determining, using the second reconstructed prediction residual value for the second channel, a second predicted residual value for a third channel of the video data.

4. The method of claim 1 , further comprising:
transforming and quantizing a second reconstructed prediction residual value for a third channel of the video data and inverse quantizing and inverse transforming the transformed and quantized second reconstructed prediction residual value for the third channel to generate a third reconstructed prediction residual value for the third channel, wherein determining the first predicted residual value for the second channel comprises determining the first predicted residual value using both the first reconstructed prediction residual value and the third reconstructed prediction residual value.

5. The method of claim 4 , wherein the first channel comprises a luma channel, the second channel comprises a chroma channel, and the third channel comprises another chroma channel.

6. The method of claim 1 , wherein determining the first predicted residual value comprises multiplying the first reconstructed prediction residual value by the first linear model parameter corresponding to the particular subset and adding the second linear model parameter corresponding to the particular subset.

7. The method of claim 1 , wherein determining the first reconstructed prediction residual value comprises transforming and quantizing a previously predicted residual value for the first channel and inverse quantizing and inverse transforming the transformed and quantized previously predicted residual value to generate the first reconstructed prediction residual value.

8. An apparatus, comprising:
a processor configured to:
divide a codec dependent range of possible reconstructed prediction residual values for a first channel of video data into a plurality of subsets of possible reconstructed prediction residual values;
assign each of the subsets a first linear model parameter and a second linear model parameter for linear cross-channel residual prediction;
determine a first reconstructed prediction residual value at a particular pixel position within a picture having a plurality of pixel positions for the first channel of video data, wherein the particular position is within a first block of the video data;
determine, using the first reconstructed prediction residual value, a first predicted residual value at the particular pixel position for a second channel of the video data, wherein to determine the first predicted residual value, the processor is configured to:
determine the first reconstructed prediction residual value at the particular pixel position for the first channel is within a particular subset of the subsets of possible reconstructed prediction residual values; and
predict the first predicted residual value for the second channel based on first and second linear model parameters corresponding to the particular subset of the subsets of possible reconstructed prediction residual values;

determine second reconstructed prediction residual values for a second block of the first channel of the video data and third reconstructed prediction residual values for a third block of the first channel of the video data, wherein the second block is intra coded and the third block is inter coded;
determine, based on the second reconstructed prediction residual values, second predicted residual values for the second channel of the video data using linear cross-channel residual prediction, wherein the linear cross-channel residual prediction is performed responsive to the second block being intra coded; and
skip cross-channel residual prediction for the second channel of the video data for the third block responsive to the third block being inter coded.

9. The apparatus of claim 8 , wherein when the first channel comprises a luma channel, the second channel comprises a chroma channel and when the first channel comprises a chroma channel, the second channel comprises one of a luma channel or another chroma channel.

10. The apparatus of claim 8 , wherein the processor to determine the first predicted residual value comprises the processor to multiply the first reconstructed prediction residual value by the first linear model parameter corresponding to the particular subset and add the second linear model parameter corresponding to the particular subset.

11. A computer-implemented method, comprising:
at a video encoder:
dividing a codec dependent range of possible reconstructed prediction residual values for a first channel of video data into a plurality of subsets of possible reconstructed prediction residual values;
assigning each of the subsets of possible reconstructed prediction residual values a first linear model parameter and a second linear model parameter for linear cross-channel residual prediction;
determining a first reconstructed prediction residual value at a particular pi

Assignees

• Xu Lidong
• Chiu Yi Jen
• Han Yu
• Zhang Wenhao
• Intel Corp

Inventors

• Xu Lidong
• Chiu Yi-Jen
• Han Yu
• Zhang Wenhao

Classifications

• H04N19/176

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

• H04N19/159

  Prediction type, e.g. intra-frame, inter-frame or bidirectional frame prediction · CPC title

• H04N19/186

  the unit being a colour or a chrominance component · CPC title

• H04N19/61

  in combination with predictive coding · CPC title

• H04N19/105 Primary

  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 title