Block and mode reordering to facilitate parallel intra prediction and motion vector prediction
View Patent ↗A method for processing a plurality of sub-blocks in a block of video is disclosed. The method generally includes the steps of (A) intra predicting a first group of the sub-blocks in a first quadrant of the block, (B) intra predicting a second group of the sub-blocks in a second quadrant of the block after starting the intra predicting of the first group and (C) intra predicting a third group of the sub-blocks in the first quadrant after starting the intra predicting of the second group, wherein the first group and the third group together account for all of the sub-blocks in the first quadrant.
1. A method for processing intra prediction of a plurality of sub-blocks in a block of video according to a plurality of prediction modes, said method comprising the steps of:
(A) intra predicting a first group of said sub-blocks in a first quadrant of said block, each sub-block using a corresponding sequence of prediction modes;
(B) intra predicting a second group of said sub-blocks in a second quadrant of said block after starting said intra predicting of said first group, each sub-block using a corresponding sequence of prediction modes;
(C) intra predicting a third group of said sub-blocks in said first quadrant after starting said intra predicting of said second group, each sub-block using a corresponding sequence of prediction modes, wherein said first group and said third group together account for all of said sub-blocks in said first quadrant;
wherein said plurality of sub-blocks in said block is processed in at least two parallel processes, and every two successively-indexed sub-blocks in said block are processed in two distinct processes in the at least two parallel processes, and wherein the sequence of prediction modes corresponding to each sub-blocks is scheduled in order of availability of context needed for the plurality prediction mode.
2. The method according to claim 1 , further comprising the steps of:
intra predicting a fourth group of said sub-blocks in said second quadrant after starting said intra predicting of said third group, each sub-block using a corresponding sequence of prediction modes;
intra predicting a fifth group of said sub-blocks in a third quadrant of said block after starting said intra predicting of said fourth group, each sub-block using a corresponding sequence of prediction modes; and
intra predicting a sixth group of sub-blocks in said second quadrant after starting said intra predicting of said fifth group, each sub-block using a corresponding sequence of prediction modes, wherein said second group, said fourth group and said sixth group together account for all of said sub-blocks in said second quadrant.
3. The method according to claim 1 , wherein a coding order of said sub-blocks comprises an index number 0, an index number 1, an index number 2, an index number 3, an index number 4, an index number 5, an index number 6, and index number 7, an index number 8, an index number 9, an index number 10, an index number 11, an index number 12, an index number 13, an index number 14 and an index number 15 sequentially from first to last.
4. The method according to claim 3 , wherein (i) said first quadrant comprises an upper-left quadrant of said block and (ii) a prediction order in said upper-left quadrant comprises said index number 0, said index number 1, said index number 2 and said index number 4 for an upper-left corner, an upper-right corner, a lower-left corner and a lower-right corner of said upper-left quadrant respectively.
5. The method according to claim 3 , wherein (i) said second quadrant comprises an upper-right quadrant of said block and (ii) said prediction order in said upper-right quadrant comprises said index number 3, said index number 5, said index number 6 and said index number 8 for an upper-left corner, an upper-right corner, a lower-left corner and a lower-right corner of said upper-right quadrant respectively.
6. The method according to claim 1 , wherein a prediction order results in
(i) at most two of said sub-blocks being dependent on an availability of a left context and
(ii) at most four of said sub-blocks being dependent on an availability of an above-right context.
7. The method according to claim 1 , wherein an encoding of said video is compliant with an International Telecommunication Union-Telecommunications Standardization Sector Recommendation H.264 accounting for said intra prediction of said sub-blocks in said step (A), said step (B) and said step (C).
8. The method according to claim 1 , further comprising: before intra-predicting the sub-blocks, determining an order of applying the intra-predicting on the indexed sub-blocks as dynamic ordering, the dynamic ordering of the indexed sub-blocks provided to minimize a number of dependencies of the indexed sub-blocks on image data values of adjacent sub-blocks.
9. The method according to claim 8 , further comprising buffering data of one or more indexed sub-blocks in order to dynamically change the ordering for intra-predicting the sub-blocks, the dependencies of the sub-blocks having variation based on a mode of intra sub-block prediction.
10. The method according to claim 9 , wherein the buffering buffers only data for only a single indexed sub-block at one time.
11. The method of claim 1 , wherein for every two successively-indexed sub-blocks, a first of the two sub-blocks is prediction processed in a first of the parallel processes for intra prediction, while a second of the two sub-blocks is prediction processed in a second of the two parallel processes for intra prediction.
12. The method of claim 1 , further comprising: before intra-predicting the sub-blocks, determining which of a plurality of prediction modes to eliminate from the intra-predicting processes based on statistics calculated for the block.
13. A method for processing intra prediction of a plurality of sub-blocks in a block of video according to a plurality of prediction modes, comprising the steps of:
(A) predicting a group of plurality of motion vectors corresponding to said sub-blocks in a first half of said block, each sub-block using a corresponding sequence of prediction modes;
(B) predicting a first one of said motion vectors corresponding to said sub-blocks in a second half of said block after starting said predicting of said group, each sub-block using a corresponding sequence of prediction modes;
(C) predicting a second one of said motion vectors corresponding to said sub-blocks in said first half after starting said predicting of said first one to said motion vector, each sub-block using a corresponding sequence of prediction modes; and
wherein said plurality of sub-blocks in said block is processed in at least two parallel processes, and every two successively-indexed sub-blocks in said block are processed in two distinct processes in the at least two parallel processes, wherein the sequence of prediction modes corresponding to each sub-block is scheduled in order of availability of context needed for the plurality of prediction modes.
14. The method according to claim 13 , further comprising the step of:
predicting a third one of said motion vectors corresponding to said sub-blocks in said second half after starting said predicting of said second one of said motion vectors.
15. The method according to claim 13 , wherein a coding order of said sub-blocks comprises an index number 0, an index number 1, an index number 2, an index number 3, an index number 4, an index number 5, an index number 6 and an index number 7 sequentially from first to last.
16. The method according to claim 15 , wherein (i) said first half comprises a left half of said block and (ii) a prediction order in an upper quadrant of said left half comprises said index number 0, said index number 1, said index number 2, and said index number 4 for an upper-left corner, an upper-right corner, a lower-left corner and a lower-right corner respectively.
17. The method according to claim 15 , wherein (i) said first half comprises a left half of said block and (ii) a prediction order in said left half comprises said index number 0, said index number 1, said index number 2 and said index number 4 for an upper-left corner, an upper-right corner, a lower-left corner and a lower-right corner respectively.
18. The method according to claim 15 , wherein (i) said first half comprises a left half of said block and (ii) a prediction order in said left half comprises said index number 0, said index number 2, said index number 4 and said index number 6 from top to bottom respectively.
19. The method according to claim 13 , wherein an encoding of said video is compliant with an International Telecommunication Union-Telecommunications Standardization Sector Recommendation H.264 accounting for said predicting of said motion vectors in said step (A), said step (B) and said step (C).