Video coding using cross-component linear model

1. A method for video decoding, comprising:

deriving a first parameter α and a second parameter β for a cross-component linear model (CCLM) mode by using a number of neighboring reconstructed luma samples and neighboring chroma samples of a coding unit (CU); and

generating a predicted chroma sample in the CU by using the first parameter α and the second parameter β;

wherein deriving the first parameter α and the second parameter β comprises: deriving the first parameter α and the second parameter β by using four neighboring down-sampled luma samples and four corresponding chroma samples, and wherein each down-sampled luma sample and its corresponding chroma sample have the same row index and the same column index; and

wherein the four neighboring down-sampled luma samples and four corresponding chroma samples include one of following sets:

the top sample of the left neighboring samples, the leftmost sample of the above neighboring samples, the bottom sample of the left neighboring samples, and the rightmost sample of the above neighboring samples;

the quarter of width of the leftmost sample of the above neighboring samples, the quarter of height of the top sample of the left neighboring samples, the bottom sample of the left neighboring samples, and the rightmost sample of the above neighboring samples;

the quarter of width of the leftmost sample of the above neighboring samples, the quarter of height of the top sample of the left neighboring samples, the three quarter of width of the leftmost sample of the above neighboring samples, the three quarter of height of the top sample of the left neighboring samples;

the one-eighth of width of the leftmost sample of the above neighboring samples, the three-eighth of width of the leftmost sample of the above neighboring samples, the five-eighth of width of the leftmost sample of the above neighboring samples, the seven-eighth of width of the leftmost sample of the above neighboring samples; and

the one-eighth of height of the top sample of the left neighboring samples, the three-eighth of height of the top sample of the left neighboring samples, the five-eighth of height of the top sample of the left neighboring samples, the seven-eighth of height of the top sample of the left neighboring samples.

2. The method of claim 1 , wherein generating the predicted chroma sample comprises:

obtaining the predicted chroma sample based on following equation:

Pred C ( x,y )=α*rec L ′( x,y )+β

wherein pred C (x,y) is the predicted chroma sample, rec L ′(x,y) is down-sampled reconstructed luma samples of the CU, x indicates a row index, and y indicates a column index.

3. The method of claim 1 , wherein deriving the first parameter α and the second parameter β comprises:

obtaining the first parameter α and the second parameter β based on following equations:

α

=

y

B

-

y

A

x

B

-

x

A

β

=

y

A

·

α

⁢

⁢

x

A

wherein x A , x B , y A and y B are derived based on weighted averages of the neighboring chroma samples and their corresponding down-sampled luma samples.

4. The method of claim 1 , wherein deriving the first parameter α and the second parameter β by using four neighboring down-sampled luma samples and four corresponding chroma samples comprises:

obtaining the first parameter α and the second parameter β by using a weighted average luma value x A of two smaller luma samples x A0 , x A1 from the four neighboring down-sampled luma samples, a weighted average luma value x B of two larger luma samples x B0 , x B1 from the four neighboring down-sampled luma samples, a weighted average chroma value y A of two chroma samples y A0 , y A1 corresponding to the two smaller luma samples, and a weighted average chroma value y B of two chroma samples y B0 , y B1 corresponding to the two larger luma samples,

wherein x A , x B , y A , and y B are obtained based on following equations:

x A =( w 1* x A0 +w 2* x A1 +offset1)>> N 1;

x B =( w 3* x B0 +w 4* x B1 +offset2)>> N 2;

y A =( w 1* y A0 +w 2* y A1 +offset1)>> N 1;

y B =( w 3* y B0 +w 4* y B1 +offset2)>> N 2;

wherein w1+w2=(1<<N1), offset1=1<<(N1-1), w3+w4=(1<<N2), offset2=1<<(N2-1),

wherein w1 is a first weight factor, w2 is a second weight factor, w3 is a third weight factor, w4 is a fourth weight factor, N1 is the first average value, N2 is the second average value, offset1 is a first offset factor, and offset2 is a second offset factor.

5. A computing device, comprising:

at least one processor;

a non-transitory storage coupled to the one or more processors; and

a plurality of programs stored in the non-transitory storage that, when executed by the processor, cause the computing device to perform acts comprising:

deriving a first parameter α and a second parameter β for a cross-component linear model (CCLM) mode by using a number of neighboring reconstructed luma samples and neighboring chroma samples of a coding unit (CU); and

generating a predicted chroma sample in the CU by using the first parameter α and the second parameter β;

wherein deriving the first parameter α and the second parameter β comprises: deriving the first parameter α and the second parameter β by using four neighboring down-sampled luma samples and four corresponding chroma samples, and wherein each down-sampled luma sample and its corresponding chroma sample have the same row index and the same column index; and

wherein the four neighboring down-sampled luma samples and four corresponding chroma samples include one of following sets:

the top sample of the left neighboring samples, the leftmost sample of the above neighboring samples, the bottom sample of the left neighboring samples, and the rightmost sample of the above neighboring samples;

the quarter of width of the leftmost sample of the above neighboring samples, the quarter of height of the top sample of the left neighboring samples, the bottom sample of the left neighboring samples, and the rightmost sample of the above neighboring samples;

the quarter of width of the leftmost sample of the above neighboring samples, the quarter of height of the top sample of the left neighboring samples, the three quarter of width of the leftmost sample of the above neighboring samples, the three quarter of height of the top sample of the left neighboring samples;

the one-eighth of width of the leftmost sample of the above neighboring samples, the three-eighth of width of the leftmost sample of the above neighboring samples, the five-eighth of width of the leftmost sample of the above neighboring samples, the seven-eighth of width of the leftmost sample of the above neighboring samples; and

the one-eighth of height of the top sample of the left neighboring samples, the three-eighth of height of the top sample of the left neighboring samples, the five-eighth of height of the top sample of the left neighboring samples, the seven-eighth of height of the top sample of the left neighboring samples.

6. The computing device of claim 5 , wherein generating the predicted chroma sample comprises:

obtaining the predicted chroma sample based on following equation:

Pred C ( x,y )=α*rec L ′( x,y )+β

wherein pred C (x,y) is the predicted chroma sample, rec L ′:(x,y) is the down-sampled reconstructed luma samples of the CU, x indicates a row index, and y indicates a column index.

7. The computing device of claim 5 , wherein deriving the first parameter α and the second parameter β comprises:

obtaining the first parameter α and the second parameter β based on following equations:

α

=

y

B

-

y

A

x

B

-

x

A

β

=

y

A

·

α

⁢

⁢

x

A

wherein x A , x B , y A and y B are derived based on weighted averages of the neighboring chroma samples and their corresponding down-sampled luma samples.

8. The computing device of claim 5 , wherein deriving the first parameter α and the second parameter β by using four neighboring down-sampled luma samples and four corresponding chroma samples comprises:

obtaining the first parameter α and the second parameter β by using a weighted average luma value x A of two smaller luma samples x A0 , x A1 from the four neighboring down-sampled luma samples, a weighted average luma value x B of two larger luma samples x B0 , x B1 from the four neighboring down-sampled luma samples, a weighted average chroma value y A of two chroma samples y A0 , y A1 corresponding to the two smaller luma samples, and a weighted average chroma value y B of two chroma samples y B0 , y B1 corresponding to the two larger luma samples,

wherein x A , x B , y A , and y B are obtained based on following equations:

x A =( w 1* x A0 w 2* x A1 +offset1)>> N 1;

x B =( w 3* x B0 +w 4* x B1 +offset2)>> N 2;

y A =( w 1* y A0 +w 2* y A1 +offset1)>> N 1;

y B =( w 3* y B0 +w 4* y B1 +offset2)>> N 2;

wherein w1+w2=(1<<N1), offset1=1<<(N1-1), w3+w4=(1<<N2), offset2=1<<(N2-1), wherein w1 is a first weight factor, w2 is a second weight factor, w3 is a third weight factor, w4 is a fourth weight factor, N1 is the first average value, N2 is the second average value, offset1 is a first offset factor, and offset2 is a second offset factor.

9. A non-transitory computer readable storage medium storing bitstream that when received by a decoding device causes the decoding device to perform acts comprising:

deriving a first parameter α and a second parameter β for a cross-component linear model (CCLM) mode by using a number of neighboring reconstructed luma samples and neighboring chroma samples of a coding unit (CU); and

generating a predicted chroma sample in the CU by using the first parameter α and the second parameter β;

wherein deriving the first parameter α and the second parameter β comprises: deriving the first parameter α and the second parameter β by using four neighboring down-sampled luma samples and four corresponding chroma samples, and wherein each down-sampled luma sample and its corresponding chroma sample have the same row index and the same column index; and

wherein the four neighboring down-sampled luma samples and four corresponding chroma samples include one of following sets:

the top sample of the left neighboring samples, the leftmost sample of the above neighboring samples, the bottom sample of the left neighboring samples, and the rightmost sample of the above neighboring samples;

the quarter of width of the leftmost sample of the above neighboring samples, the quarter of height of the top sample of the left neighboring samples, the bottom sample of the left neighboring samples, and the rightmost sample of the above neighboring samples;

the quarter of width of the leftmost sample of the above neighboring samples, the quarter of height of the top sample of the left neighboring samples, the three quarter of width of the leftmost sample of the above neighboring samples, the three quarter of height of the top sample of the left neighboring samples;

the one-eighth of width of the leftmost sample of the above neighboring samples, the three-eighth of width of the leftmost sample of the above neighboring samples, the five-eighth of width of the leftmost sample of the above neighboring samples, the seven-eighth of width of the leftmost sample of the above neighboring samples; and

the one-eighth of height of the top sample of the left neighboring samples, the three-eighth of height of the top sample of the left neighboring samples, the five-eighth of height of the top sample of the left neighboring samples, the seven-eighth of height of the top sample of the left neighboring samples.

10. The non-transitory computer readable storage medium of claim 9 , wherein generating the predicted chroma sample comprises:

obtaining the predicted chroma sample based on following equation:

Pred C ( x,y )=α*rec L ′( x,y )+β

wherein pred C (x,y) is the predicted chroma sample, rec L ′(x,y) is down-sampled reconstructed luma samples of the CU, x indicates a row index, and y indicates a column index.

11. The non-transitory computer readable storage medium of claim 9 , wherein deriving the first parameter α and the second parameter β comprises:

obtaining the first parameter α and the second parameter β based on following equations:

α

=

y

B

-

y

A

x

B

-

x

A

β

=

y

A

·

α

⁢

⁢

x

A

wherein x A , x B , y A and y B are derived based on weighted averages of the neighboring chroma samples and their corresponding down-sampled luma samples.

12. The non-transitory computer readable storage medium of claim 9 , wherein deriving the first parameter α and the second parameter β by using four neighboring down-sampled luma samples and four corresponding chroma samples comprises:

obtaining the first parameter α and the second parameter β by using a weighted average luma value x A of two smaller luma samples x A0 , x A1 from the four neighboring down-sampled luma samples, a weighted average luma value x B of two larger luma samples x B0 , x B1 from the four neighboring down-sampled luma samples, a weighted average chroma value y A of two chroma samples y A0 , y A1 corresponding to the two smaller luma samples, and a weighted average chroma value y B of two chroma samples y B0 , y B1 corresponding to the two larger luma samples,

wherein x A , x B , y A , and y B are obtained based on following equations:

x A =( w 1* x A0 +w 2* x A1 +offset1)>> N 1;

x B =( w 3* x B0 +w 4* x B1 +offset2)>> N 2;

y A =( w 1* y A0 +w 2* y A1 +offset1)>> N 1;

y B =( w 3* y B0 +w 4* y B1 +offset2)>> N 2;

wherein w1+w2=(1<<N1), offset1=1<<(N1-1), w3+w4=(1<<N2), offset2=1<<(N2-1), wherein w1 is a first weight factor, w2 is a second weight factor, w3 is a third weight factor, w4 is a fourth weight factor, N1 is the first average value, N2 is the second average value, offset1 is a first offset factor, and offset2 is a second offset factor.