Contextual design for encoding and decoding syntax elements in interpredictive mode

Abstract

The present invention provides a method and system for determining the context for encoding and decoding syntax elements of a video stream in interpredictive mode. [Solution] A method for decoding a current prediction block within a current coding block of a video stream includes the steps of: receiving a flag indicating that the current prediction block is interpredicted; deriving a first parameter for quantifying the extent to which the current prediction block shares a reference frame with a spatially adjacent prediction block of the current coding block; deriving a second parameter associated with the use of motion vector prediction within a spatially adjacent prediction block of the current coding block; and, depending on the first and second parameters, selecting a coding context for at least one of a set of syntaxes associated with interprediction of the current prediction block.

Claims

1. A method for decoding the current predicted block within the current encoded block of a video stream, The process of receiving a flag that indicates the current prediction block should be interpreted, The process involves deriving a first parameter for quantifying the extent to which the current prediction block shares a reference frame with a spatially adjacent prediction block of the current encoding block, A step of deriving a second parameter associated with the use of motion vector prediction within the spatially adjacent prediction block of the current coding block, A step of selecting an encoding context for at least one of a set of syntaxes associated with the interprediction of the current prediction block, depending on the first and second parameters, wherein the step of selecting the encoding context is independent of or further based on any derivation of any time-motion vector prediction (TMVP) candidate block, and the set of syntaxes includes at least a dynamic motion vector reference list (DRL) index and a single reference interprediction mode index. Methods that include...

2. The first parameter, represented by "nearest_match", includes a predefined set of binary row reference frame matching indices that indicate whether there is a reference frame match between the current prediction block and at least one spatially adjacent prediction block in the row of the prediction block immediately above the current coding block, and binary column reference frame matching indices that indicate whether there is a reference frame match between the current prediction block and at least one spatially adjacent prediction block in the column of the prediction block immediately to the left of the current coding, The second parameter, represented by "newmv_count", indicates the number of prediction blocks in the spatially adjacent prediction blocks of the current prediction block, which has at least one motion vector predicted by at least one reference motion vector. The method according to claim 1.

3. A step of deriving a first context index represented by "newmv_ctx" from the first parameter and the second parameter, by setting newmv_ctx = nearest_match + (newmv_count > 0), wherein the () operator is configured to generate 1 and 0 when the logical expression enclosed by the () operator is true and false, respectively. The steps include: selecting a set of coding contexts for the DRL index of the current prediction block, depending on the first and second parameters, at least via newmv_ctx; The method according to claim 2, further comprising:

4. The method according to claim 3, wherein the selection of the set of coding contexts for the DRL index of the current prediction block, depending on the first parameter and the second parameter, is via maximum newmv_ctx.

5. A step of deriving a second context index represented by "refmv_ctx" from the first parameter and the second parameter, where A is a predefined or signaled positive integer, by setting refmv_ctx = nearest_match + (newmv_count < A), The current prediction block's interprediction is determined to be in single-reference interprediction mode. In response to the determination, the steps include selecting an encoding context for the single-reference interprediction mode of the current prediction block, depending on the first and second parameters, via the maximum newmv_ctx and refmv_ctx. The method according to claim 4, further comprising:

6. A step of deriving a second context index represented by "refmv_ctx" from the first parameter and the second parameter, where A is a predefined or signaled positive integer, by setting refmv_ctx = nearest_match + Boolean(newmv_count < A), In response to the determination that the interprediction of the current prediction block is in a composite reference interprediction mode, the steps include selecting an encoding context for the composite reference interprediction mode of the current prediction block via MAX newmv_ctx and REFMV_ctx, depending on the first and second parameters, The method according to claim 3, further comprising:

7. The method according to claim 3, further comprising the step of selecting an encoding context for the single-reference inter-prediction mode of the current prediction block, depending on the first and second parameters, via MAX NEWMV_CTX, in response to the determination that the inter-prediction of the current prediction block is a single-reference inter-prediction mode.

8. The method according to claim 3, further comprising the step of selecting an encoding context for the composite reference inter-prediction mode of the current prediction block, depending on the first and second parameters, via MAX NEWMV_CTX, in response to the determination that the inter-prediction of the current prediction block is a composite reference inter-prediction mode.

9. The method according to claim 3, further comprising the step of selecting an encoding context for the optical flow mode of the current prediction block, depending on the first and second parameters, via MAX NEWMV_CTX, in response to the determination that an optical flow mode is to be used for the current prediction block.

10. The method according to claim 2, wherein the derivation of the first and second parameters comprises the step of scanning the M nearest leftmost columns and nearest uppermost rows of spatially adjacent prediction blocks of the current coding block in a predefined scan order, where M is a predefined or signaled positive integer.

11. The method according to claim 10, wherein the predefined scan order of the nearest left column and nearest top row of a spatially adjacent prediction block follows (L-1, B) → (R, T-1) → (L-1, T) → (L, T-1) → (L-1, B-1) → (L+1, T-1) → (L-1, C_row) → (C_column, T-1), where L, R, T, and B represent the block coordinates of the left column, right column, top row, and bottom row of the current coding block, respectively, and C_row and C_column represent the block coordinate positions of the current row and current column of the current prediction block, respectively.

12. The method according to claim 10, wherein M = 2.

13. The method according to claim 10, wherein M = 4.

14. The predefined scan order of the nearest left column and nearest top row of the spatially adjacent prediction block follows the order (R, T-1) → (L-1, B) → (L, T-1) → (L-1, T), where L, R, T, and B are the left column, right column, and first row of the current coding block. The method according to claim 13, which represents the block coordinates of the top row and the bottom row, respectively.

15. A step of deriving a first context index represented by "newmv_ctx" from the first parameter and the second parameter, by setting newmv_ctx = nearest_match + (newmv_count > 0), wherein the () operator is configured to generate 1 and 0 when the logical expression enclosed by the () operator is true and false, respectively. A step of deriving a second context index represented by "refmv_ctx" from the first parameter and the second parameter, where A is a predefined or signaled positive integer, and It further includes, The step of selecting the encoding context for at least one of the set of syntaxes includes the step of selecting the context according to newmv_ctx and refmv_ctx, The first and second parameters for deriving newmv_ctx are obtained by scanning up to L neighboring prediction blocks in the column immediately to the left and the row immediately above the current coding block. The first and second parameters for deriving refmv_ctx are obtained by scanning up to K neighboring prediction blocks in the column immediately to the left and the row immediately above the current coding block. L and K are predefined or signaled positive integers not greater than M, which is a predefined or signaled positive integer. The method according to claim 2.

16. A method for processing the current prediction block within the current encoding block of a video stream, The process of receiving a flag that indicates the current prediction block should be interpreted, A step of deriving context parameters based on checking the availability of one or more time motion vector prediction (TMVP) candidate blocks and comparing a global motion vector with available TMVP candidate blocks, wherein at least one of the one or more TMVP candidate blocks is spatially located at a position other than the top left of the current encoding block; A step of determining the coding context for the syntax element associated with the interprediction of the current prediction block, according to the context parameter. Methods that include...

17. The method according to claim 16, wherein the one or more TMVP candidate blocks include a single TMVP candidate block spatially positioned at the bottom right or center of the current encoding block.

18. The one or more TMVP candidate blocks include a plurality of TMVP candidate blocks spatially positioned at the bottom right and center positions of the current encoded block, The multiple TMVP candidate blocks are checked for availability, and the first available TMVP candidate block is compared to the global motion vector to derive the context parameters. The method according to claim 16.

19. The aforementioned one or more TMVP candidate blocks are located in the upper left corner of the current encoded block. It includes multiple TMVP candidate blocks spatially positioned at the following locations: the bottom right position and the center position. The multiple TMVP candidate blocks are checked for availability, and the first available TMVP candidate block is compared to the global motion vector to derive the context parameters. The method according to claim 16.

20. A program that causes at least one processor to perform the method according to any one of claims 1 to 19.

21. An electronic device for processing the current prediction block within the current encoding block of a video stream, The current prediction block receives a flag indicating that it should be interpreted, To derive a first parameter for quantifying the extent to which the current prediction block shares a reference frame with a spatially adjacent prediction block of the current coding block, To derive a second parameter associated with the use of motion vector prediction within the spatially adjacent prediction block of the current coding block, Depending on the first and second parameters, the selection of an encoding context for at least one of the sets of syntax associated with the interprediction of the current prediction block, wherein the selection of the encoding context is independent of or based on any derivation of any time-motion vector prediction (TMVP) candidate block, and the set of syntax includes at least a dynamic motion vector reference list (DRL) index and a single reference interprediction mode index. An electronic device comprising a memory for storing instructions for causing the electronic device to perform the aforementioned action, and a processor for executing the aforementioned instructions.

Images