Decoder-Side Motion Vector Refinement for Faster Video Reconstruction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Decoder-side motion vector refinement techniques in video coding face processing delays due to how motion vectors are computed and reconstructed, leading to inefficiencies and potential blocking artifacts.
Innovation Solution
Implementing decoder-side predictor refinement techniques that utilize two sets of motion vectors, where one set is used for parsing and motion vector derivation, and the other for reconstruction, allowing early reconstruction of unrefined motion vectors and using refined motion vectors for post-parsing processing to avoid additional blocking artifacts.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If decoder-side motion vector refinement techniques are used to improve coding efficiency, then coding gain is improved, but processing delays increase and execution speed decreases
Solution Approach 1:
The decoding process is segmented into two distinct phases: a parsing phase using unrefined motion vectors for syntax decoding and candidate derivation, and a reconstruction phase using refined motion vectors for motion compensation. This segmentation allows each phase to use the most appropriate motion vector type, improving overall execution efficiency while maintaining coding gain.
Solution Approach 2:
Unrefined motion vectors are reconstructed early during the parsing phase to enable syntax decoding and candidate list derivation before refinement is complete. This preliminary reconstruction allows the decoding pipeline to proceed without waiting for the computationally intensive refinement process to finish.
2Productivity
If unrefined motion vectors are used for parsing and derivation, then execution speed is improved, but blocking artifacts may occur in post-parsing processing
Solution Approach 1:
Different motion vector qualities are applied to different processing stages: unrefined motion vectors are used for parsing and candidate derivation where high precision is not critical, while refined motion vectors are used for motion compensation and post-parsing processing where high precision is essential to avoid blocking artifacts. This local quality differentiation optimizes both speed and quality.
3Manufacturing precision
If refined motion vectors are used for all decoding processes, then manufacturing precision is improved, but device complexity and processing time increase
Solution Approach 1:
The system dynamically selects which motion vectors to refine based on the decoding stage and processing requirements. Not all motion vectors undergo refinement - only those needed for post-parsing processing. This dynamic approach reduces unnecessary computational complexity while maintaining precision where it matters most.
Data Source
AI summary
Methods, apparatus, and computer readable media configured to determine motion vectors for encoders and decoders are disclosed. A decoder receives compressed video data related to a set of frames. The decoder calculates, using a decoder-side predictor refinement technique, a new motion vector for a current frame from the set of frames, wherein the new motion vector estimates motion for the current frame based on one or more reference frames. The calculation includes retrieving a first motion vector associated with the current frame, executing a first portion of the decoding process using the first motion vector, retrieving a second motion vector associated with the current frame that is different than the first motion vector, and executing a second portion of the decoding process using the second motion vector.