Motion Vector Memory Reduction in Video Encoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video compression technologies, such as HEVC, face significant motion vector memory requirements due to the need to store multiple motion vectors and their predictors, leading to increased storage needs and potential coding efficiency penalties.
Innovation Solution
The method involves selecting one motion vector from pairs in a reference frame and mapping all other motion vectors to this selected frame, eliminating the need to store reference frame indices and reducing memory size by using dummy motion vectors and scaling, thereby compressing motion vector information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If multiple motion vectors are stored in reference frames to improve prediction accuracy, then coding efficiency is improved, but motion vector memory size increases
Solution Approach 1:
The patent extracts only the necessary motion vector information by selecting one motion vector from pairs in reference frames. This selective extraction reduces the quantity of stored motion vectors while maintaining essential prediction capabilities, directly addressing the contradiction between storage size and prediction accuracy.
Solution Approach 2:
The patent applies scaling to motion vectors to map them between different reference frames. By changing the parameter representation of motion vectors through scaling operations, the system can reuse fewer stored vectors across multiple reference frames, reducing memory requirements while preserving prediction functionality.
2Quantity of substance
If motion vector memory size is reduced to decrease storage requirements, then device complexity is reduced, but coding efficiency may be penalized
Solution Approach 1:
The patent merges the functionality of multiple reference frames by mapping motion vectors from different frames to a single selected reference frame. This consolidation allows the system to store fewer motion vectors while still supporting multiple reference frame operations, reducing memory size without significantly impacting coding efficiency.
Solution Approach 2:
The selected reference frame and its motion vectors serve multiple purposes by acting as a universal reference for multiple original reference frames. This multi-functionality allows the system to reduce the number of stored motion vectors while maintaining the ability to perform predictions against multiple reference frames.
3Quantity of substance
If motion vectors are mapped to a single reference frame to reduce storage, then motion vector memory size is reduced, but reference frame index information is lost
Solution Approach 1:
The patent creates a copy of the selected reference frame and maps motion vectors from other reference frames to this copy through scaling operations. This copying approach allows the system to eliminate the need to store separate motion vectors for each reference frame, reducing memory size while the scaling factors preserve the essential spatial relationships.
Data Source
AI summary
Motion vectors of a first reference frame are permitted to point to a plurality of further reference frames. A method of storing the motion vectors comprises, when a block of the first reference frame has two motion vectors (V2A, V2B) initially, selecting one of the two motion vectors, the non-selected motion vector not being stored. The selected motion vector may be scaled. This can reduce the motion vector memory size.