Granted Patent B2

US 12,587,656 · App. 18/697,628 · Granted Mar 24, 2026

Inventors: Seunghwan Kim (Seoul, KR); Jie Zhao (Seoul, KR); Seethal Paluri (Seoul, KR)

Assignee: LG Electronics Inc.

H04N19/159H04N19/105H04N19/147H04N19/176H04N19/70

Abstract

Claims (61)

1 . An image decoding method performed by a decoding apparatus, the image decoding method comprising:

obtaining image information from a bitstream;

determining whether at least one of Template-based Intra Mode Derivation (TIMD) mode or Decoder Side Intra Mode Derivation (DIMD) mode is applied to a current block;

generating a prediction sample of the current block based on at least one of the TIMD mode or the DIMD mode; and

generating a reconstructed sample of the current block based on the prediction sample of the current block,

wherein the TIMD mode is derived based on a template of the current block,

wherein the DIMD mode is derived based on a gradient of neighboring reference samples of the current block,

wherein the image decoding method further comprises constructing a most probable mode (MPM) list including a candidate intra prediction mode,

wherein an intra prediction mode is derived based on the MPM list,

wherein the MPM list includes a primary MPM list and a secondary MPM list, and

wherein, based on the TIMD mode being applied to the current block, the candidate intra prediction mode derived based on the TIMD mode is inserted into the MPM list.

2 . The image decoding method of claim 1 , wherein the prediction sample is generated based on performing weighted averaging or blending based on a first prediction mode and a second prediction mode,

wherein the first prediction mode is derived based on the TIMD mode, and

wherein the second prediction mode is derived based on the DIMD mode.

3 . The image decoding method of claim 2 , wherein a weight for the weighted averaging is determined based on rate distortion cost.

4 . The image decoding method of claim 2 , wherein the prediction sample is generated based on performing the weighted averaging or the blending further based on a third prediction mode,

wherein the third prediction mode is derived as a planner mode, and

wherein the second prediction mode is different from the planar mode.

5 . The image decoding method of claim 2 , wherein the image information includes a syntax element indicating a fusion mode or an intra merge mode for the weighted averaging or the blending.

6 . The image decoding method of claim 5 , wherein the syntax element is based on a context coded model or fixed length coding.

7 . The image decoding method of claim 1 , wherein the image information includes a sequence parameter set (SPS) and a picture parameter set (PPS),

wherein information on whether the TIMD mode is applied to the current block is included in at least one of the SPS or the PPS, and

wherein information on whether the DIMD mode is applied to the current block is included in at least one of the SPS or the PPS.

8 . The image decoding method of claim 1 , wherein an intra prediction mode for the TIMD mode is derived as at least one of intra modes of the current block, and

wherein a maximum number of the intra modes of the current block is 131.

9 . The image decoding method of claim 1 , wherein an intra prediction mode for the DIMD mode is derived as at least one of intra modes of the current block, and

wherein a maximum number of the intra modes of the current block is 131.

10 . The image decoding method of claim 1 , further comprising constructing a most probable mode (MPM) list including a candidate intra prediction mode,

wherein the intra prediction mode is derived based on the MPM list,

wherein the MPM list includes a primary MPM list and a secondary MPM list, and

wherein, based on the DIMD mode being applied to the current block, the intra prediction mode derived based on the DIMD mode is inserted into the MPM list.

11 . An image encoding method performed by a encoding apparatus, the image encoding method comprising:

determining whether at least one of Template-based Intra Mode Derivation (TIMD) mode or Decoder Side Intra Mode Derivation (DIMD) mode is applied to a current block;

generating a prediction sample of the current block based on at least one of the TIMD mode or the DIMD mode;

generating prediction-related information based on the prediction sample of the current block; and

encoding image information including the prediction-related information,

wherein the TIMD mode is derived based on a template of the current block,

wherein the DIMD mode is derived based on a gradient of neighboring reference samples of the current block,

wherein the image decoding method further comprises constructing a most probable mode (MPM) list including a candidate intra prediction mode,

wherein an intra prediction mode is derived based on the MPM list,

wherein the MPM list includes a primary MPM list and a secondary MPM list, and

wherein, based on the TIMD mode being applied to the current block, the candidate intra prediction mode derived based on the TIMD mode is inserted into the MPM list.

12 . The image encoding method of claim 11 , wherein the prediction sample is generated based on performing weighted averaging or blending based on a first prediction mode and a second prediction mode,

wherein the first prediction mode is derived based on the TIMD mode, and

wherein the second prediction mode is derived based on the DIMD mode.

13 . The image encoding method of claim 12 , wherein a weight for the weighted averaging is determined based on rate distortion cost.

14 . The image encoding method of claim 12 , wherein the prediction sample is generated based on performing the weighted averaging or the blending further based on a third prediction mode,

wherein the third prediction mode is derived as a planner mode, and

wherein the second prediction mode is different from the planar mode.

15 . The image encoding method of claim 12 , wherein the image information includes a syntax element indicating a fusion mode or an intra merge mode for the weighted averaging or the blending.

16 . The image encoding method of claim 11 , wherein an intra prediction mode for the TIMD mode is derived as at least one of intra modes of the current block, and

wherein a maximum number of the intra modes of the current block is 131.

17 . A method of transmitting data for image, the method comprising:

obtaining a bitstream for the image, wherein the bitstream is generated based on determining whether at least one of Template-based Intra Mode Derivation (TIMD) mode or Decoder Side Intra Mode Derivation (DIMD) mode is applied to a current block, generating a prediction sample of the current block based on at least one of the TIMD mode or the DIMD mode, generating prediction-related information based on the prediction sample of the current block, and encoding image information including the prediction-related information; and

transmitting the data including the bitstream,

wherein the TIMD mode is derived based on a template of the current block,

wherein the DIMD mode is derived based on a gradient of neighboring reference samples of the current block,

wherein the image decoding method further comprises constructing a most probable mode MPM) list including a candidate intra prediction mode,

wherein an intra prediction mode is derived based on the MPM list,

wherein the MPM list includes a primary MPM list and a secondary MPM list, and

wherein, based on the TIMD mode being applied to the current block, the candidate intra prediction mode derived based on the TIMD mode is inserted into the MPM list.

Assignments (1)

Continuity (2)

References Cited (34)