Luma-based chroma intra-prediction for video coding

1. A method comprising:

receiving, by at least one processor, a picture;

dividing, by the at least one processor, the picture into a plurality of coding units (CUs) arranged in rows including a first row and a second row, wherein the first row includes a first CU associated with a line buffer configured to store reconstructed luma samples;

performing sample padding of at least one of the reconstructed luma samples to obtain padded luma samples;

down sampling, by the at least one processor, a first reconstructed luma block of the first CU, to generate a down sampled first reconstructed luma block;

computing, by the at least one processor, parameters α and β of a linear model based on a top neighboring reconstructed luma sample and a left neighboring reconstructed luma sample of the first reconstructed luma block and reconstructed neighboring chroma samples of a chroma block corresponding to the first reconstructed luma block, and wherein the computing is configured to use the padded luma samples; and

computing, by the at least one processor, samples of a first predicted chroma block from corresponding samples of the down sampled first reconstructed luma block using the parameters α and β.

2. The method of claim 1 , wherein the first reconstructed luma block is in the first row.

3. The method of claim 1 , wherein the first reconstructed luma block is not in the first row.

4. The method of claim 1 , wherein the left neighboring reconstructed luma sample is immediately left of the first reconstructed luma block.

5. The method of claim 2 , further comprising:

down sampling, by the at least one processor, a second reconstructed luma block of the first CU, wherein the second reconstructed luma block is not at a top border of the first CU, to generate a down sampled second reconstructed luma block;

computing, by the at least one processor, parameters α and β of the linear model by linear regression using immediate top neighboring reconstructed luma samples, top neighboring reconstructed luma samples immediately above the immediate top neighboring reconstructed luma samples, and left neighboring reconstructed luma samples of the second reconstructed luma block and reconstructed neighboring chroma samples of a chroma block corresponding to the second reconstructed luma block; and

computing, by the at least one processor, samples of a second predicted chroma block from corresponding samples of the down sampled second reconstructed luma block using the linear model and the parameters.

6. An encoder comprising:

at least one processor; and

a non-transitory computer readable storage medium storing a program for execution by the at least one processor, the program including instructions to:

receive a picture;

divide the picture into a plurality of coding units (CUs) arranged in rows including a first row and a second row, wherein the first row includes a first CU associated with a line buffer configured to store reconstructed luma samples;

perform sample padding of at least one of the reconstructed luma samples to obtain padded luma samples;

down sample a first reconstructed luma block of the first CU, to generate a down sampled first reconstructed luma block;

compute parameters α and β of a linear model based on a top neighboring reconstructed luma sample and a left neighboring reconstructed luma sample of the first reconstructed luma block and reconstructed neighboring chroma samples of a chroma block corresponding to the first reconstructed luma block, and wherein the instructions to compute the parameters α and β are configured to use reconstructed luma samples in the first portion of the second row, to use the padded luma samples; and

compute samples of a first predicted chroma block from corresponding samples of the down sampled first reconstructed luma block using the parameters α and β.

7. The encoder of claim 6 , wherein the first reconstructed luma block is in the first row.

8. The encoder of claim 6 , wherein the first reconstructed luma block is not in the first row.

9. The encoder of claim 6 , wherein the left neighboring reconstructed luma sample is immediately left of the first reconstructed luma block.

10. The encoder of claim 7 , wherein the instructions further comprise instructions to:

down sample a second reconstructed luma block of the first CU, wherein the second reconstructed luma block is not at a top border of the first CU, to generate a down sampled second reconstructed luma block;

compute parameters α and β of the linear model by linear regression using immediate top neighboring reconstructed luma samples, top neighboring reconstructed luma samples immediately above the immediate top neighboring reconstructed luma samples, and left neighboring reconstructed luma samples of the second reconstructed luma block and reconstructed neighboring chroma samples of a chroma block corresponding to the second reconstructed luma block; and

compute samples of a second predicted chroma block from corresponding samples of the down sampled second reconstructed luma block using the linear model and the parameters.

11. The encoder of claim 6 , wherein the plurality of CUs is a plurality of largest coding units (LCUs) and the first CU is an LCU.

12. The encoder of claim 6 , wherein the instructions further comprise instructions to:

encode the picture using the first predicted chroma block, to generate an encoded picture; and

transmit, by a transmitter, the encoded picture.

13. The method of claim 1 , wherein the plurality of CUs is a plurality of largest coding units (LCUs) and the first CU is an LCU.

14. The method of claim 1 , further comprising:

encoding the picture using the first predicted chroma block, to generate an encoded picture; and

transmitting, by a transmitter, the encoded picture.

15. A non-transitory computer readable storage medium storing a program for execution by at least one processor, the program including instructions to:

receive a picture;

divide the picture into a plurality of coding units (CUs) arranged in rows including a first row and a second row, wherein the first row includes a first CU associated with a line buffer configured to store reconstructed luma samples;

perform sample padding of at least one of the reconstructed luma samples to obtain padded luma samples;

down sample a first reconstructed luma block of the first CU to generate a down sampled first reconstructed luma block;

compute parameters α and β of a linear model based on a top neighboring reconstructed luma sample and left neighboring reconstructed luma sample[s] of the first reconstructed luma block and reconstructed neighboring chroma samples of a chroma block corresponding to the first reconstructed luma block, and wherein the instructions to compute the parameters α and β use the padded luma samples; and

compute samples of a first predicted chroma block from corresponding samples of the down sampled first reconstructed luma block using the parameters.

16. The non-transitory computer readable storage medium of claim 15 , wherein the plurality of CUs is a plurality of largest coding units (LCUs) and the first CU is an LCU.

17. The non-transitory computer readable storage medium of claim 15 , wherein the instructions further comprise instructions to:

encode the picture using the first predicted chroma block, to generate an encoded picture; and

transmit, by a transmitter, the encoded picture.