Granted Patent B2

US 12,407,839 · App. 18/750,699 · Granted Sep 2, 2025

Inventors: Yi-wen Chen (San Diego, CA); Xianglin Wang (San Diego, CA)

Assignee: BEIJING DAJIA INTERNET INFORMATION TECHNOLOGY CO., LTD.

H04N19/186H04N19/105H04N19/132H04N19/176H04N19/30H04N19/44H04N19/59

Abstract

Claims (46)

1. A method for decoding a video signal, comprising:

reconstructing a luma block corresponding to a chroma block, wherein the luma block is adjacent to a plurality of reconstructed neighboring luma samples, and wherein the chroma block is adjacent to a plurality of reconstructed neighboring chroma samples;

computing a plurality of down-sampled luma samples from the plurality of reconstructed neighboring luma samples;

identifying, from a sub-group of the plurality of computed down-sampled luma samples, two down-sampled maximum luma samples, wherein the two down-sampled maximum luma samples correspond to two first reconstructed chroma samples of the plurality of reconstructed neighboring chroma samples respectively, and the sub-group is composed of a predefined number of the computed down-sampled luma samples among the plurality of computed down-sampled luma samples;

identifying, from the sub-group of the plurality of computed down-sampled luma samples, two down-sampled minimum luma samples, wherein the two down-sampled minimum luma samples correspond to two second reconstructed chroma samples of the plurality of reconstructed neighboring chroma samples respectively;

averaging the two down-sampled maximum luma samples, the two down-sampled minimum luma samples, the two first reconstructed chroma samples, and the two second reconstructed chroma samples, respectively, to obtain an averaged down-sampled maximum luma sample, an averaged down-sampled minimum luma sample, an averaged first reconstructed chroma sample and an averaged second reconstructed chroma sample;

fitting a linear model through the averaged down-sampled maximum luma sample, the averaged down-sampled minimum luma sample, the averaged first reconstructed chroma sample, and the averaged second reconstructed chroma sample;

computing down-sampled luma samples from luma samples of the reconstructed luma block, wherein each down-sampled luma sample corresponds to a chroma sample of the chroma block; and

predicting chroma samples of the chroma block by applying the linear model to the corresponding computed down-sampled luma samples.

2. The method of claim 1 , wherein the chroma block and the luma block are encoded using a 4:2:0 or 4:2:2 chroma sub-sampling scheme, and wherein the chroma block and the luma block have different resolutions.

3. The method of claim 1 , wherein the plurality of reconstructed neighboring luma samples includes luma samples located above the reconstructed luma block and/or luma samples to left of the reconstructed luma block.

4. The method of claim 1 , wherein the computing down-sampled luma samples from luma samples of the reconstructed luma block comprises performing a weighted average of six neighboring luma samples to the luma sample.

5. The method of claim 1 , wherein the fitting the linear model comprises fitting a linear equation through a data point associated with the averaged down-sampled maximum luma sample and the averaged first reconstructed chroma sample and a data point associated with the averaged down-sampled minimum luma sample and the averaged second reconstructed chroma sample.

6. A decoding device comprising:

one or more processors;

memory coupled to the one or more processors; and

a plurality of programs stored in the memory that, when executed by the one or more processors, cause the one or more processors to perform acts comprising:

reconstructing a luma block corresponding to a chroma block, wherein the luma block is adjacent to a plurality of reconstructed neighboring luma samples, and wherein the chroma block is adjacent to a plurality of reconstructed neighboring chroma samples;

computing a plurality of down-sampled luma samples from the plurality of reconstructed neighboring luma samples;

identifying, from a sub-group of the plurality of computed down-sampled luma samples, two down-sampled maximum luma samples, wherein the two down-sampled maximum luma samples correspond to two first reconstructed chroma samples of the plurality of reconstructed neighboring chroma samples respectively, and the sub-group is composed of a predefined number of the computed down-sampled luma samples among the plurality of computed down-sampled luma samples;

identifying, from the sub-group of the plurality of computed down-sampled luma samples, two down-sampled minimum luma samples, wherein the two down-sampled minimum luma samples correspond to two second reconstructed chroma samples of the plurality of reconstructed neighboring chroma samples respectively;

averaging the two down-sampled maximum luma samples, the two down-sampled minimum luma samples, the two first reconstructed chroma samples, and the two second reconstructed chroma samples, respectively, to obtain an averaged down-sampled maximum luma sample, an averaged down-sampled minimum luma sample, an averaged first reconstructed chroma sample and an averaged second reconstructed chroma sample;

fitting a linear model through the averaged down-sampled maximum luma sample, the averaged down-sampled minimum luma sample, the averaged first reconstructed chroma sample, and the averaged second reconstructed chroma sample;

computing down-sampled luma samples from luma samples of the reconstructed luma block, wherein each down-sampled luma sample corresponds to a chroma sample of the chroma block; and

predicting chroma samples of the chroma block by applying the linear model to the corresponding computed down-sampled luma samples.

7. The decoding device of claim 6 , wherein the chroma block and the luma block are encoded using a 4:2:0 or 4:2:2 chroma sub-sampling scheme, and wherein the chroma block and the luma block have different resolutions.

8. The decoding device of claim 6 , wherein the plurality of reconstructed neighboring luma samples includes luma samples located above the reconstructed luma block and/or luma samples to left of the reconstructed luma block.

9. The decoding device of claim 6 , wherein the computing down-sampled luma samples from luma samples of the reconstructed luma block comprises performing a weighted average of six neighboring luma samples to the luma sample.

10. The decoding device of claim 6 , wherein the fitting the linear model comprises fitting a linear equation through a data point associated with the averaged down-sampled maximum luma sample and the averaged first reconstructed chroma sample and a data point associated with the averaged down-sampled minimum luma sample and the averaged second reconstructed chroma sample.

11. A method for storing a bitstream, comprising:

performing an encoding method to generate a bitstream; and

storing the bitstream,

wherein the encoding method comprises:

reconstructing a luma block corresponding to a chroma block, wherein the luma block is adjacent to a plurality of reconstructed neighboring luma samples, and wherein the chroma block is adjacent to a plurality of reconstructed neighboring chroma samples;

computing a plurality of down-sampled luma samples from the plurality of reconstructed neighboring luma samples;

identifying, from a sub-group of the plurality of computed down-sampled luma samples, two down-sampled maximum luma samples, wherein the two down-sampled maximum luma samples correspond to two first reconstructed chroma samples of the plurality of reconstructed neighboring chroma samples respectively, and the sub-group is composed of a predefined number of the computed down-sampled luma samples among the plurality of computed down-sampled luma samples;

identifying, from the sub-group of the plurality of computed down-sampled luma samples, two down-sampled minimum luma samples, wherein the two down-sampled minimum luma samples correspond to two second reconstructed chroma samples of the plurality of reconstructed neighboring chroma samples respectively;

averaging the two down-sampled maximum luma samples, the two down-sampled minimum luma samples, the two first reconstructed chroma samples, and the two second reconstructed chroma samples, respectively, to obtain an averaged down-sampled maximum luma sample, an averaged down-sampled minimum luma sample, an averaged first reconstructed chroma sample and an averaged second reconstructed chroma sample;

fitting a linear model through the averaged down-sampled maximum luma sample, the averaged down-sampled minimum luma sample, the averaged first reconstructed chroma sample, and the averaged second reconstructed chroma sample;

computing down-sampled luma samples from luma samples of the reconstructed luma block, wherein each down-sampled luma sample corresponds to a chroma sample of the chroma block; and

predicting chroma samples of the chroma block by applying the linear model to the corresponding computed down-sampled luma samples,

wherein the bitstream is to be decoded by the method of claim 1 .

12. The method of claim 11 , wherein the chroma block and the luma block are encoded using a 4:2:0 or 4:2:2 chroma sub-sampling scheme, and wherein the chroma block and the luma block have different resolutions.

13. The method of claim 11 , wherein the plurality of reconstructed neighboring luma samples includes luma samples located above the reconstructed luma block and/or luma samples to left of the reconstructed luma block.

14. The method of claim 11 , wherein the computing down-sampled luma samples from luma samples of the reconstructed luma block comprises performing a weighted average of six neighboring luma samples to the luma sample.

15. The method of claim 11 , wherein the fitting the linear model comprises fitting a linear equation through a data point associated with the averaged down-sampled maximum luma sample and the averaged first reconstructed chroma sample and a data point associated with the averaged down-sampled minimum luma sample and the averaged second reconstructed chroma sample.

Assignments (1)

Continuity (7)

References Cited (75)