Motion Compensation Using Estimated Sub-Pixel Matching Costs
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The complexity of sub-pixel search in motion compensation methods for video coding is high due to the large number of sub-pixels that need to be searched, leading to increased computational complexity and performance loss in current methods.
Innovation Solution
The method estimates approximate matching costs between a to-be-processed block and sub-pixel blocks within a search range based on the matching costs of integer pixel blocks, allowing for motion compensation without the need to search all sub-pixels, thereby reducing search complexity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If sub-pixel search is performed to improve motion compensation accuracy, then matching precision is improved, but computational complexity increases significantly
Solution Approach 1:
The patent divides the search process into two stages: first searching integer pixels to identify candidate positions, then only searching sub-pixels at these candidate positions. This segmentation of the search space reduces the total number of sub-pixels that need to be evaluated while still achieving accurate motion compensation at the optimal position.
Solution Approach 2:
The patent performs preliminary integer pixel search before sub-pixel search to identify candidate positions. This preliminary action narrows down the search range, so that subsequent sub-pixel search is only performed at a limited number of candidate positions rather than across the entire search range, reducing computational complexity.
2Measurement precision
If full sub-pixel search is performed to improve matching accuracy, then motion estimation precision is improved, but processing time increases
Solution Approach 1:
The search process is segmented into integer pixel search followed by targeted sub-pixel search at candidate positions only. This segmentation avoids the need to evaluate all sub-pixels across the entire search range, significantly reducing processing time while maintaining motion estimation precision at the optimal position.
Solution Approach 2:
Instead of performing complete sub-pixel search across the entire search range, the patent performs partial sub-pixel search only at candidate positions identified by integer pixel search. This partial action is sufficient to achieve accurate motion estimation without the excessive time cost of a complete search.
3Reliability
If the search range includes many sub-pixels to ensure accurate matching, then matching reliability is improved, but computational load increases
Solution Approach 1:
The patent segments the search process into two phases: integer pixel search to identify candidate positions, and then sub-pixel search limited to these candidate positions. This segmentation ensures that reliable matching is achieved at the optimal position while avoiding the computational load of searching all sub-pixels in the range.
Solution Approach 2:
Integer pixel search is performed as a preliminary step to identify candidate positions before conducting sub-pixel search. This preliminary action ensures that sub-pixel search is only performed where needed, maintaining matching reliability at candidate positions while reducing overall computational load.
Data Source
AI summary
A motion compensation method, includes: acquiring a plurality of matching costs between a to-be-processed block and a plurality of first pixel blocks, the plurality of first pixel blocks corresponding to a plurality of integer pixels within a search range in a reference frame; based on the plurality of matching costs, estimating matching costs between the to-be-processed block and a plurality of second pixel blocks to obtain a plurality of approximate matching costs, the plurality of second pixel blocks corresponding to a plurality of sub-pixels within the search range in the reference frame; and performing motion compensation on the to-be-processed block according to the plurality of matching costs and the plurality of approximate matching costs.