Motion Estimation Method Using Global Motion Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current motion estimation algorithms are inefficient due to their focus on local motion, leading to increased computational requirements when dealing with global motion in video compression, such as camera or object movement, which limits their effectiveness in multimedia communication systems.
Innovation Solution
A method that classifies motion vectors into direction classes and determines global motion types to set a preferential search area within the full-search method, reducing unnecessary calculations by prioritizing areas where the motion vector is likely to be found, thereby enhancing the efficiency of motion estimation in video encoding.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the full-search algorithm searches all pixel positions within the specified search area to find the motion vector, then the accuracy of motion estimation is improved, but the computational load increases significantly
Solution Approach 1:
The patent applies preliminary action by performing a coarse search first to identify a preferential search area, then conducting a fine search only within that reduced area. This two-stage approach prepares the search space in advance, eliminating unnecessary calculations while preserving the ability to find accurate motion vectors when needed.
Solution Approach 2:
The patent implements local quality by applying different search strategies to different regions: a coarse search is performed on the entire search area to identify regions with significant motion, while a fine search is performed only on the identified preferential areas. This localized differentiation optimizes computational resources based on actual motion characteristics in each region.
2Productivity
If the search area is limited to a small region around the original point to reduce calculations, then the encoding speed is improved, but the ability to capture global motion is reduced
Solution Approach 1:
The patent applies dynamics by making the search area adaptive rather than fixed. The preferential search area is dynamically determined based on the actual motion characteristics detected in the current frame, allowing the algorithm to expand the search area when global motion is present and contract it when local motion dominates, thus optimizing both speed and adaptability.
Solution Approach 2:
The patent implements segmentation by dividing the search process into two distinct stages: a coarse search stage that covers the entire search area to detect global motion patterns, and a fine search stage that focuses on specific preferential areas. This segmentation allows the algorithm to handle both local and global motion effectively while maintaining encoding efficiency.
3Measurement precision
If the hierarchical variable size block matching algorithm is used to improve motion estimation accuracy, then the motion vector precision is improved, but the computational complexity increases sharply
Solution Approach 1:
The patent applies partial action by performing the computationally intensive fine search only on preferential areas that are determined to contain the actual motion vectors, rather than performing exhaustive searches on the entire search area. This partial application of the detailed search algorithm significantly reduces computational complexity while maintaining motion vector precision.
Solution Approach 2:
The patent uses preliminary coarse search to identify preferential areas before performing the detailed hierarchical variable size block matching. This preliminary action filters out regions that do not contain significant motion, allowing the complex algorithm to be applied only where necessary, thus reducing overall computational complexity while preserving accuracy.
Data Source
Figure 1
Figure 2
Figure 3~4
AI summary
A motion estimation method for enhancing a video compression speed, and a video encoder using the same are provided. The motion estimation method includes determining a global motion type of a previous frame from motion vectors of the previous frame, setting a search area for a specified motion block included in a current frame based on the global motion type, and searching for a motion vector within the set search area.