Blending method for inter prediction with geometry partition

1. A method of video processing, comprising:

determining, for a conversion between a block of a video and a bitstream of the video, weights to use for a weighted average of prediction samples along an edge of two partitions of the block, wherein the two partitions are generated according to a geometry partition mode, wherein the weights are selected from a first weighting factor group of at least two weighting factor groups based on a first motion vector associated with a first partition and a second motion vector associated with a second partition; and

performing the conversion based on the determining,

wherein the first motion vector associated with the first partition is denoted as Mv0(x0, y0), the second motion vector associated with the second partition is denoted as Mv1(x1, y1), and wherein the second motion vector is scaled as Mv1′(x1′, y1′) according to a reference picture that the first motion vector points to, and

wherein the weights in the first weighting factor group are used in a case that a value of a function F with respect to (x1′, x0, y1′, y0) is larger than or equal to a threshold, and

wherein the threshold varies for different blocks having different characteristics.

2. The method of claim 1 , wherein the weights are selected from the first weighting factor group in a case that at least one reference picture of the block is located in a current picture that includes the block, in a case that all motion vectors of the two partitions have an integer precision, or in a case the block is a chroma block.

3. The method of claim 1 , wherein the threshold is predefined or the threshold is signaled in a sequence parameter set (SPS), a video parameter set (VPS), a picture parameter set (PPS), a picture header, a slice header, a tile group header, a coding tree unit (CTU), or a region.

4. The method of claim 3 , wherein in a case where the threshold is predefined, the threshold is 16<<MV_PRECISION, wherein << represents a left-shift operation, and wherein MV_PRECISION represents a precision of a stored motion vector.

5. The method of claim 4 , wherein MV_PRECISION is set to 4 in case the precision of the stored motion vector is 1/16-pel.

6. The method of claim 1 , wherein the characteristics include a coding mode, a dimension, a slice type, a tile type, or a picture type.

7. The method of claim 1 , wherein F is max(abs(x1′−x0), abs (y1′−y0)), min(abs(x1′−x0), abs (y1′−y0)), abs(x1′−x0)+abs (y1′−y0) or an average of Mv0 and Mv1′.

8. The method of claim 1 , wherein two reference pictures of the two partitions are same, and wherein the two reference pictures are different than a current reference picture in which the block is located.

9. The method of claim 1 , wherein the weights are selected from the first weighting factor group based on a characteristic associated with one or more reference pictures of the block.

10. The method of claim 9 , wherein the characteristic comprises at least one of a picture-order-count (POC) distance between two reference pictures of the block, and a picture-order-count (POC) distance between a reference picture of the block and a current picture in which the block is located.

11. The method of claim 1 , wherein the block is split into multiple partitions based on at least a splitting pattern.

12. The method of claim 1 , wherein the geometry partition mode comprises a triangular partitioning mode.

13. The method of claim 1 , wherein at least one of the multiple two partitions partitioned in the geometry partition mode is a non-square and non-rectangular block.

14. The method of claim 1 , wherein the conversion includes encoding the block of the video into the bitstream.

15. The method of claim 1 , wherein the conversion includes decoding the block of the video from the bitstream.

16. An apparatus for processing video data comprising a processor and a non-transitory memory with instructions thereon, wherein the instructions upon execution by the processor, cause the processor to:

determine, for a conversion between a block of a video and a bitstream of the video, weights to use for a weighted average of prediction samples along an edge of two partitions of the block, wherein the two partitions are generated according to a geometry partition mode, wherein the weights are selected from a first weighting factor group of at least two weighting factor groups based on a first motion vector associated with a first partition and a second motion vector associated with a second partition; and

perform the conversion based on the determining,

wherein the first motion vector associated with the first partition is denoted as Mv0(x0, y0), the second motion vector associated with the second partition is denoted as Mv1(x1, y1), and wherein the second motion vector is scaled as Mv1′(x1′, y1′) according to a reference picture that the first motion vector points to, and

wherein the weights in the first weighting factor group are used in a case a value of a function F with respect to (x1′, x0, y1′, y0) is larger than or equal to a threshold, and wherein the threshold varies for different blocks having different characteristics.

17. The apparatus of claim 16 , wherein the characteristics include a coding mode, a dimension, a slice type, a tile type, or a picture type.

18. The apparatus of claim 16 , wherein F is max(abs(x1′−x0), abs (y1′−y0)), min(abs(x1′−x0), abs (y1′−y0)), abs(x1′−x0)+abs (y1′−y0) or an average of Mv0 and Mv1′.

19. The apparatus of claim 16 , wherein two reference pictures of the two partitions are same, and wherein the two reference pictures are different than a current reference picture in which the block is located.

20. A method for storing a bitstream of a video, wherein the method comprises:

determining, for generating the bitstream from a block of a video, weights to use for a weighted average of prediction samples along an edge of two partitions of the block, wherein the two partitions are generated according to a geometry partition mode, wherein the weights are selected from a first weighting factor group of at least two weighting factor groups based on a first motion vector associated with a first partition and a second motion vector associated with a second partition;

generating the bitstream based on the determining, and

storing the bitstream in a non-transitory computer-readable recording medium,

wherein a characteristic of the block comprises a precision of a motion vector of the block, or an indicator indicating a latency attribute of the block,

wherein the first motion vector associated with the first partition is denoted as Mv0(x0, y0), the second motion vector associated with the second partition is denoted as Mv1(x1, y1), and wherein the second motion vector is scaled as Mv1′(x1′, y1′) according to a reference picture that the first motion vector points to, and

wherein the weights in the first weighting factor group are used in a case a value of a function F with respect to (x1′, x0, y1′, y0) is larger than or equal to a threshold, and

wherein the threshold varies for different blocks having different characteristics.