Motion Estimation Method Using Grouped Block Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional motion estimation methods require significant calculation resources and time due to the complexity of comparing each possible searching block with a matching window, especially as the searching area increases, while alternative methods that accelerate the process often compromise on accuracy.
Innovation Solution
A motion estimation method that calculates differences between a single searching block and a matching window to obtain total differences for multiple searching blocks simultaneously, allowing for a full search while reducing calculation time and memory resources by theoretical N times.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional full search algorithm compares each searching block with matching window one after another, then accurate motion vector matching is achieved, but calculation time and computation resources increase significantly
Solution Approach 1:
The patent divides the searching area into multiple groups, where each group contains multiple searching blocks. Instead of comparing each block individually, the method compares one representative searching block from each group with the matching window to obtain N total differences simultaneously. This segmentation reduces the number of comparisons from M*N individual block comparisons to M group comparisons, where N is the number of blocks per group, thereby reducing calculation time while maintaining full search accuracy.
2Reliability
If conventional full search algorithm compares each searching block with matching window, then complete search coverage is achieved, but memory buffer requirements increase
Solution Approach 1:
The patent segments the searching area into multiple groups with each group containing N searching blocks. By selecting one representative block from each group and calculating N total differences simultaneously, the method reduces the number of buffers needed to store matching results. Instead of requiring buffers for all M*N individual blocks, the system only needs buffers for M group representatives, reducing memory requirements by a factor of N while maintaining complete search coverage through the group-based comparison approach.
3Measurement precision
If searching area size increases to improve matching coverage, then motion estimation accuracy is improved, but calculation complexity increases
Solution Approach 1:
The patent applies segmentation by dividing the searching area into multiple groups, where each group contains N searching blocks. This allows the system to handle larger searching areas more efficiently by processing groups rather than individual blocks. The complexity is reduced from O(M*N) individual comparisons to O(M) group comparisons, where M is the number of groups and N is the number of blocks per group. This maintains comprehensive search coverage across large areas while reducing calculation complexity through the group-based parallel difference calculation approach.
Data Source
AI summary
A motion estimation method applied to a video signal having a first frame and a second frame is provided. The motion estimation method includes: capturing a matching window from the first frame; capturing a searching area from the second frame, the searching area including a plurality of searching blocks each having a size equal to that of the matching window; selecting one of the searching blocks; calculating a difference between pixel data of the selected searching blocks and pixel data of the matching window, so as to obtain N total differences corresponding to N searching blocks; and determining a motion vector corresponding to the matching window according to the N total differences.