SYSTEMS AND METHOD FOR INTER PREDICTION BASED ON A MERGE MODE

Abstract

The present disclosure relates to systems and methods for inter prediction based on a Merge mode. The methods may include determining whether a current block in an image frame satisfies a division condition, and in response to determining that the current block satisfies the division condition, obtaining two sub-blocks by dividing the current block according to a division mode. The division mode may include a rectangular division. The methods may further include determine, based on the two sub-blocks, a prediction result of the current block.

Description

Claims

1. A system for inter prediction based on a Merge mode, comprising: at least one storage medium including a set of instructions; andat least one processor in communication with the at least one storage medium, wherein when executing the set of instructions, the at least one processor is directed to cause the system to: determine whether a current block in an image frame satisfies a division condition;in response to determining that the current block satisfies the division condition, obtain two sub-blocks by dividing the current block according to a division mode, wherein the division mode includes a rectangular division; anddetermine, based on the two sub-blocks, a prediction result of the current block.

2. The system of claim 1, wherein the rectangular division is converted from an inclined division.

3. The system of claim 2, wherein to obtain the two sub-blocks by dividing the current block according to a division mode, the at least one processor is directed to cause the system to: obtain two intermediate sub-blocks by dividing the current block according to the inclined division;determine a count of pixels of at least one of the two intermediate sub-blocks; andobtain the two sub-blocks by converting the inclined division of the current block into the rectangular division of the current block based on the count of pixels.

4. The system of claim 2, wherein to obtain the two sub-blocks by dividing the current block according to a division mode, the at least one processor is directed to cause the system to: obtain two intermediate sub-blocks by dividing the current block according to the inclined division;determine a dividing point between the two intermediate sub-blocks; andobtain the two sub-blocks by converting the inclined division of the current block into the rectangular division of the current block based on the dividing point between the two intermediate sub-blocks.

5. The system of claim 2, wherein to determine, based on the two sub-blocks, a prediction result of the current block, the at least one processor is directed to cause the system further to: obtain motion compensation information of the two sub-blocks by performing, based on the rectangular division, a motion compensation on the two sub-blocks;determine, based on the motion compensation information and the inclined division, target prediction values of the two sub-blocks; anddetermine, based on the target prediction values of the two sub-blocks, the prediction result of the current block.

6. The system of claim 2, wherein to determine, based on the two sub-blocks, a prediction result of the current block, the at least one processor is directed to cause the system further to: obtain two predicted sub-blocks corresponding to the two sub-blocks;obtain filtered pixel values by performing a first filtering operation on pixel values at a junction region of the two predicted sub-blocks;designate the filtered pixel values as target prediction values of the two predicted sub-blocks; anddetermine, based on the target prediction values of the two sub-blocks, the prediction result of the current block.

7. The system of claim 6, wherein to obtain filtered pixel values of the two predicted sub-blocks by performing a first filtering operation on pixel values at a junction of the two predicted sub-blocks, the at least one processor is directed to cause the system to: filter, based on a tap filter, the pixel values at the junction of the two predicted sub-blocks, wherein coefficients of the tap filter include {1, 2, 1}.

8. The system of claim 6, wherein to obtain two predicted sub-blocks corresponding to the two sub-blocks, the at least one processor is directed to cause the system to: obtain motion compensation information of the two sub-blocks by performing, based on the rectangular division, a motion compensation on the two sub-blocks;determine, based on the motion compensation information and the inclined division, target prediction values of the two sub-blocks;obtain two updated sub-blocks corresponding to the two sub-blocks by performing, based on the target prediction values of the two sub-blocks, a second filter operation on the two sub-blocks; anddesignate the two updated sub-blocks as the two predicted sub-blocks.

9. The system of claim 1, wherein the division mode includes at least one of a horizontal division, a vertical division, or a diagonal division.

10. The system of claim 1, wherein the division mode is adjustable according to operations including: adjusting an effective length of a reference weight matrix in the division mode; andadjusting the division mode based on the adjusted effective length.

11. The system of claim 1, wherein offset directions of the division mode include at least one of: up, down, left, right, upper left, upper right, lower left or lower right.

12. The system of claim 1, wherein offset distances of the division mode include at least one of: ¼ pixel accuracy, ½ pixel accuracy, 1 pixel accuracy, 2 times of pixel accuracy, 4 times of pixel accuracy, 8 times of pixel accuracy, 16 times of pixel accuracy, or 32 times of pixel accuracy.

13. The system of claim 1, wherein the system is applied to the third generation audio video coding standard (AVS3).

14. The system of claim 1, wherein the division mode includes a plurality of division manners, each of the plurality of division manners corresponds to two candidate sub-blocks, and to obtain two sub-blocks by dividing the current block according to a division mode, the at least one processor is directed to cause the system to: for each two candidate sub-blocks corresponding to one of the plurality of division manners, obtain motion compensation information of the two candidate sub-blocks byperforming, based on the rectangular division, a motion compensation on the two candidate sub-blocks;determine target motion compensation information in the motion compensation information corresponding to the plurality of division manners; anddesignate two candidate sub-blocks corresponding to the target motion compensation information as the two sub-blocks.

15. The system of claim 14, wherein the plurality of division manners correspond to a plurality of step sizes, the plurality of step sizes relating to division differences between the plurality of division manners.

16. A method for inter prediction based on a Merge mode, implemented on a computing device including at least one processor and at least one storage, comprising: determining whether a current block in an image frame satisfies a division condition;in response to determining that the current block satisfies the division condition, obtaining two sub-blocks by dividing the current block according to a division mode, wherein the division mode includes a rectangular division; anddetermining, based on the two sub-blocks, a prediction result of the current block.

17. The method of claim 16, wherein the rectangular division is converted from an inclined division.

18-25. (canceled)

26. The method of claim 16, wherein offset directions of the division mode include at least one of: up, down, left, right, upper left, upper right, lower left or lower right.

27. The method of claim 16, wherein offset distances of the division mode include at least one of: ¼ pixel accuracy, ½ pixel accuracy, 1 pixel accuracy, 2 times of pixel accuracy, 4 times of pixel accuracy, 8 times of pixel accuracy, 16 times of pixel accuracy, or 32 times of pixel accuracy.

28-45. (canceled)

46. A non-transitory computer readable medium, comprising executable instructions that, when executed by at least one processor, direct the at least one processor to perform a method, the method comprising: determining whether a current block in an image frame satisfies a division condition;in response to determining that the current block satisfies the division condition, obtaining two sub-blocks by dividing the current block according to a division mode, wherein the division mode includes a rectangular division; anddetermining, based on the two sub-blocks, a prediction result of the current block.