Method and apparatus for designing low frequency non-separable transform
According to the disclosure of the present disclosure: an LFNST set index for a current block may be derived on the basis of an intra prediction mode on the basis of the intra prediction mode for the current block being a WAIP mode, the LFNST set index may be derived as predetermined one value; and on the basis of the intra prediction mode for the current block being an NAIP mode, the LFNST set index may be derived as one of predetermined N values. Accordingly, it is possible to maximize the coding performance that can be brought about by LFNST within the implementation complexity allowed by a next standard.
1 . An image decoding method performed by a decoding apparatus, the method comprising:
obtaining image information through a bitstream, wherein the image information includes residual information, information related to an intra prediction mode, and low-frequency non-separable transform (LFNST) index information;
deriving transform coefficients for a current block from the residual information;
deriving an LFNST set index for the current block based on an intra prediction mode for the current block derived from the information related to the intra prediction mode;
deriving an LFNST matrix for the current block based on an LFNST index derived from the LFNST index information and the LFNST set index;
deriving modified transform coefficients based on the transform coefficients and the LFNST matrix;
generating residual samples for the current block based on the modified transform coefficients; and
generating a reconstructed picture based on the residual samples,
wherein based on the intra prediction mode for the current block being a wide angle intra prediction (WAIP) mode, the LFNST set index is derived as one predetermined value,
wherein based on the intra prediction mode for the current block being a normal angle intra prediction (NAIP) mode, the LFNST set index is derived as one of N predetermined values, and
wherein the LFNST matrix for the current block with both of a width and a height being equal to or greater than 16 includes R×96 matrix and R is one of 16, 32, 48, 64, 80 and 96.
2 . The image decoding method of claim 1 ,
wherein N is 35, and
wherein one LFNST set index derived based on the intra prediction mode of the current block being the WAIP mode is equal to one of 35 LFNST set indexes derived based on the intra prediction mode of the current block being the NAIP mode.
3 . The image decoding method of claim 2 ,
wherein the LFNST set index for the current block is derived based on the following table, and
Intra pred. mode
−14
−13
−12
−11
−10
−9
−8
−7
−6
−5
−4
−3
−2
−1
0
1
LFNST set index
2
2
2
2
2
2
2
2
2
2
2
2
2
2
0
1
Intra pred. mode
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
LFNST set index
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
Intra pred. mode
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
LFNST set index
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
Intra pred. mode
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
LFNST set index
34
33
32
31
30
29
28
27
26
25
24
23
22
21
20
19
Intra pred. mode
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
LFNST set index
18
17
16
15
14
13
12
11
10
9
8
7
6
5
4
3
Intra pred. mode
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
LFNST set index
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
wherein Intra pred. mode represents the intra prediction mode for the current block, and the case of that the value of the Intra pred. mode is one of −14 to −1 and 67 to 80 represents the intra prediction mode for the current block being the WAIP mode.
4 . The image decoding method of claim 1 ,
wherein N is 35, and
wherein one LFNST set index derived based on the intra prediction mode of the current block being the WAIP mode is different from 35 LFNST set indexes derived based on the intra prediction mode of the current block being the NAIP mode.
5 . The image decoding method of claim 4 ,
wherein the LFNST set index for the current block is derived based on the following table, and
Intra pred. mode
−14
−13
−12
−11
−10
−9
−8
−7
−6
−5
−4
−3
−2
−1
0
1
LFNST set index
35
35
35
35
35
35
35
35
35
35
35
35
35
35
0
1
Intra pred. mode
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
LFNST set index
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
Intra pred. mode
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
LFNST set index
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
Intra pred. mode
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
LFNST set index
34
33
32
31
30
29
28
27
26
25
24
23
22
21
20
19
Intra pred. mode
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
LFNST set index
18
17
16
15
14
13
12
11
10
9
8
7
6
5
4
3
Intra pred. mode
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
LFNST set index
2
35
35
35
35
35
35
35
35
35
35
35
35
35
35
wherein Intra pred. mode represents the intra prediction mode for the current block, and the case of that the value of the Intra pred. mode is one of −14 to −1 and 67 to 80 represents the intra prediction mode for the current block being the WAIP mode.
6 . The image decoding method of claim 1 ,
wherein N is 15, and
wherein one LFNST set index derived based on the intra prediction mode of the current block being the WAIP mode is equal to one of 15 LFNST set indexes derived based on the intra prediction mode of the current block being the NAIP mode.
7 . The image decoding method of claim 1 ,
wherein N is 15, and
wherein one LFNST set index derived based on the intra prediction mode of the current block being the WAIP mode is different from 15 LFNST set indexes derived based on the intra prediction mode of the current block being the NAIP mode.
8 . The image decoding method of claim 1 ,
wherein the image information further includes LFNST set list index information,
wherein the image decoding method further comprises:
deriving an LFNST set list for the current block based on an LFNST set list index for the current block derived from the LFNST set list index information, and
wherein the LFNST matrix for the current block is derived based on the LFNST index and the LFNST set index from the LFNST set list for the current block.
9 . The image decoding method of claim 8 ,
wherein the LFNST set list index is derived as a different value according to a range of a quantization parameter value, and the range of quantization parameter value is divided based on a predetermined threshold value.
10 . The image decoding method of claim 8 ,
wherein the LFNST set list index is derived as a different value according to a range of a transform block size.
11 . The image decoding method of claim 8 ,
wherein the LFNST set list index is derived as a different value according to whether a specific coding tool is applied to the current block, and
wherein the specific coding tool is at least one of a Position Dependent Intra Prediction (PDPC), matrix weighted intra prediction (MIP) mode, intra sub-partitions (ISP) mode, Adaptive Color Transform (ACT) or Multiple Transform Selection (MTS).
12 . The image decoding method of claim 8 ,
wherein the LFNST set list index is derived as a different value according to a color format of the current block.
13 . An image encoding method performed by an encoding apparatus, the method comprising:
determining an intra prediction mode for a current block, and determining a low-frequency non-separable transform (LFNST) set index based on the intra prediction mode for the current block;
generating information related to the intra prediction mode;
generating prediction samples for the current block based on the intra prediction mode;
generating residual samples for the current block based on the prediction samples;
deriving transform coefficients based on the residual samples;
deriving modified transform coefficients based on the transform coefficients and an LFNST matrix;
generating residual information based on the modified transform coefficients;
generating LFNST index information representing the LFNST matrix in an LFNST set represented by the LFNST set index; and
encoding image information including information related to the intra prediction mode, the residual information, and the LFNST index information,
wherein based on the intra prediction mode for the current block being a wide angle intra prediction (WAIP) mode, the LFNST set index is derived as one predetermined value, and
wherein based on the intra prediction mode for the current block being a normal angle intra prediction (NAIP) mode, the LFNST set index is derived as one of N predetermined values, and
wherein the LFNST matrix for the current block with both of a width and a height being equal to or greater than 16 includes R×96 matrix and R is one of 16, 32, 48, 64, 80 and 96.
14 . The image encoding method of claim 13 ,
wherein N is 35, and
wherein one LFNST set index derived based on the intra prediction mode of the current block being the WAIP mode is equal to one of 35 LFNST set indexes derived based on the intra prediction mode of the current block being the NAIP mode.
15 . A transmission method of data for an image, the method comprising:
obtaining a bitstream for the image, wherein the bitstream is generated based on determining an intra prediction mode for a current block, and determining an low-frequency non-separable transform (LFNST) set index based on the intra prediction mode for the current block, generating information related to the intra prediction mode, generating prediction samples for the current block based on the intra prediction mode, generating residual samples for the current block based on the prediction samples, deriving transform coefficients based on the residual samples, deriving modified transform coefficients based on the transform coefficients and an LFNST matrix, generating residual information based on the modified transform coefficients, generating LFNST index information representing the LFNST matrix in an LFNST set represented by the LFNST set index, and encoding image information including information related to the intra prediction mode, the residual information, and the LFNST index information; and
transmitting the data comprising the bitstream,
wherein based on the intra prediction mode for the current block being a wide angle intra prediction (WAIP) mode, the LFNST set index is derived as one predetermined value, and
wherein based on the intra prediction mode for the current block being a normal angle intra prediction (NAIP) mode, the LFNST set index is derived as one of N predetermined values, and
wherein the LFNST matrix for the current block with both of a width and a height being equal to or greater than 16 includes Rx96 matrix and R is one of 16, 32, 48, 64, 80 and 96.