Motion Estimation Search Window Prediction for Embedded Video
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video compression methods require high memory bandwidth for motion estimation, leading to increased hardware costs and power consumption due to the need for extensive data transfer and processing.
Innovation Solution
A method utilizing a square search algorithm and data reuse scheme to adaptively adjust the search window based on neighboring block data, reducing memory usage and bandwidth requirements by predicting the search window and loading only necessary data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a full search window is loaded into internal memory for motion estimation, then motion estimation accuracy is improved, but memory bandwidth requirement increases
Solution Approach 1:
The patent applies preliminary action by predicting the motion vector of the current block before performing motion estimation. The predicted motion vector is calculated using motion vectors of neighboring blocks, which allows the system to pre-determine the search window position and size. This preliminary prediction enables selective loading of only the necessary search window data into internal memory, rather than loading the entire reference frame, thus reducing memory bandwidth requirements while maintaining motion estimation accuracy.
2Measurement precision
If a large search window is used for motion estimation, then motion estimation accuracy is improved, but internal memory usage increases
Solution Approach 1:
The patent uses preliminary action by calculating the predicted motion vector before motion estimation based on neighboring block motion vectors. This prediction allows the system to determine the optimal search window size and position in advance, ensuring that only the necessary portion of the reference frame is loaded into internal memory. This approach maintains motion estimation accuracy by focusing computational resources on the most relevant search area while minimizing internal memory usage.
Solution Approach 2:
The patent applies local quality by adaptively adjusting the search window size based on the predicted motion vector and local characteristics of the current block. Instead of using a uniform large search window for all blocks, the system dynamically determines the appropriate search window dimensions for each block based on its specific requirements. This localized approach ensures that each block receives adequate search coverage for accurate motion estimation while optimizing overall internal memory usage across the entire frame.
3Loss of information
If extensive data transfer is performed for motion estimation, then motion estimation completeness is improved, but power consumption increases
Solution Approach 1:
The patent applies preliminary action by predicting motion vectors before performing motion estimation. This prediction step allows the system to identify and load only the essential data required for accurate motion estimation, rather than transferring extensive data from the reference frame. By pre-determining the search window parameters based on neighboring block motion vectors, the system minimizes data transfer operations and associated power consumption while maintaining motion estimation completeness for the critical search area.
4Measurement precision
If a full reference frame is loaded for motion estimation, then motion estimation accuracy is improved, but processing time increases
Solution Approach 1:
The patent applies preliminary action by calculating predicted motion vectors using neighboring block information before performing the actual motion estimation. This prediction step allows the system to pre-identify the search window position and size, enabling selective loading of only the necessary reference frame data into internal memory. By focusing computation on a predetermined, optimized search area rather than processing the entire reference frame, the system maintains motion estimation accuracy while significantly reducing processing time.
Solution Approach 2:
The patent applies the extraction principle by isolating and processing only the relevant search window data that is necessary for motion estimation. Instead of loading and processing the entire reference frame, the system extracts and loads only the specific region defined by the predicted search window into internal memory. This selective extraction approach maintains motion estimation accuracy by focusing computational resources on the most relevant data while reducing processing time by eliminating unnecessary computations on unrelated frame data.
Data Source
AI summary
A method for performing motion estimation is provided. The method includes the following steps: selecting a current block in a current frame; obtaining motion vectors and residual data of a plurality of neighboring blocks adjacent to the current block; setting a predetermined threshold value according to the residual data of the plurality of neighboring blocks; comparing the current block with an initial reference block in a reference frame to obtain an initial comparison result, and comparing the predetermined threshold value with the initial comparison result; determining a predicted motion vector of the current block according to the motion vectors of the plurality of neighboring blocks if the initial comparison result is larger than the predetermined threshold value; and performing a block-matching operation in a search window corresponding to the predicted motion vector to determine a corresponding reference block which best matches the current block.