Motion Estimation Using Collocated Blocks for Video Encoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional motion estimation techniques face challenges in accurately generating motion vectors for all blocks of pixels in a sequence of frames, particularly in handling occlusions and requiring high memory and processing resources, which leads to inefficiencies and inaccuracies in video encoding and frame rate conversion.
Innovation Solution
A data processing system and method that generate motion vector candidates by identifying matching pixels between frames and forming output vectors that are substantially parallel to the candidates, with the option to create zigzag vectors in occlusion regions using pixel data from the previous and next frames, reducing the need for additional frames and minimizing processing overhead.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional motion estimation techniques are used to generate motion vectors for all blocks, then motion estimation accuracy is improved, but memory and processing requirements increase significantly
Solution Approach 1:
The patent segments the motion estimation process into two distinct parts: (1) generating motion vector candidates using traditional block-matching methods, and (2) selecting the best candidate using a cost function that considers both match quality and vector consistency with neighboring blocks. This segmentation allows the system to maintain accuracy while reducing the search space and computational complexity by avoiding exhaustive searches for every block.
Solution Approach 2:
The patent applies different processing strategies to different regions of the video frame. In occluded regions where traditional motion estimation fails, the system uses a fallback mechanism that inherits motion vectors from non-occluded neighboring blocks. This local adaptation ensures accurate motion estimation in critical regions without unnecessarily increasing processing requirements for the entire frame.
2Quantity of substance
If traditional block matching is used for all blocks including occluded regions, then motion vector coverage is improved, but measurement accuracy deteriorates due to occlusion errors
Solution Approach 1:
The patent implements a dynamic motion estimation system that adapts its strategy based on local frame conditions. The cost function dynamically adjusts the weight given to match quality versus vector consistency with neighbors, and the system dynamically switches between different motion estimation modes (full search, reduced search, or neighbor inheritance) depending on whether a block is identified as occluded or not.
Solution Approach 2:
The patent introduces an intermediary cost function that acts as a mediator between the conflicting requirements of covering all blocks and maintaining accuracy in occluded regions. This cost function evaluates multiple candidate vectors and selects the one that best balances match quality with consistency to neighboring blocks, effectively filtering out erroneous vectors from occluded regions while maintaining comprehensive coverage.
3Measurement precision
If additional frames are used to resolve occlusion issues, then motion estimation accuracy is improved, but processing time and complexity increase
Solution Approach 1:
The patent enables the motion estimation system to self-correct errors in occluded regions by using information from neighboring blocks within the same frame pair. Instead of requiring additional frames or complex temporal analysis, the system uses the spatial redundancy and consistency of motion vectors across neighboring blocks to identify and correct erroneous vectors, achieving high accuracy using only the currently processed frame data.
Data Source
AI summary
A data processing system for performing motion estimation in a sequence of frames having first and second frames each divided into respective sets of blocks of pixels, includes a vector generator configured to form motion vector candidates representing mappings of pixels between the first and second frames; and a vector processor configured to, for a search block of the first frame, identify a first motion vector candidate ending in a block of the second frame collocated with the search block and form an output vector for the search block which is substantially parallel to the first motion vector candidate and represents a mapping of pixels from the search block to the second frame.