Motion Vector Refinement for Higher-Throughput Image Decoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional image encoding/decoding methods face limitations in throughput improvement due to dependency on neighboring blocks, which hinders efficient processing of high-resolution and high-quality images.
Innovation Solution
The method and apparatus minimize dependency on neighboring blocks by performing motion information refinement on current blocks before neighboring blocks are encoded/decoded, using second motion information derived from partitioning the current block into subblocks and refining motion information on these subblocks.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If motion estimation and refinement, and intra-prediction on a current block are performed after encoding/decoding on a neighboring block has completed, then manufacturing precision is improved, but productivity deteriorates
Solution Approach 1:
The patent performs motion information refinement on the current block before completing encoding/decoding of neighboring blocks. Specifically, the refinement process uses motion vectors from reference blocks that are already decoded, allowing the current block's motion information to be refined in advance without waiting for all neighboring blocks to be fully processed. This preliminary action enables parallel processing and improves throughput while maintaining encoding precision.
2Reliability
If encoding/decoding on a current block waits for neighboring block completion, then reliability is improved, but productivity deteriorates
Solution Approach 1:
The patent segments the encoding/decoding process into independent stages: neighboring block encoding/decoding, motion information extraction from reference blocks, and current block motion information refinement. These stages can be executed in parallel or overlapping time periods, reducing the critical path and improving throughput while maintaining reliability through proper dependency management.
Solution Approach 2:
The patent performs motion information refinement on the current block before neighboring block encoding/decoding completes by using motion vectors from already-decoded reference blocks. This preliminary refinement action reduces waiting time and improves throughput while maintaining encoding reliability through proper data dependency management.
3Manufacturing precision
If motion information refinement is performed on subblocks of the current block, then manufacturing precision is improved, but device complexity increases
Solution Approach 1:
The patent divides the current block into multiple subblocks and performs motion information refinement on each subblock independently. This segmentation allows for more precise motion estimation at the subblock level, improving overall encoding precision. The independent processing of subblocks also enables parallel computation, which can mitigate the increased processing complexity through efficient resource utilization.
Solution Approach 2:
The patent applies motion information refinement selectively to different subblocks of the current block, allowing each subblock to have optimized motion compensation tailored to its local characteristics. This local quality approach improves precision where needed while managing overall processing complexity through selective refinement rather than uniform processing of the entire block.
Data Source
AI summary
There is provided an image encoding/decoding method and apparatus. The image decoding method comprises decoding information on whether or not to perform motion information refinement from a bitstream, obtaining second motion information for first motion information on a current block on the basis of the information and reconstructing at least one of the current block and a neighboring block of the current block by using at least one of the first motion information and the second motion information.