Reference Area Transfer Pre-Analysis for Video Memory Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In video processing, there is a need to minimize data transfer from external memory to internal memory for motion estimation and motion compensation while optimizing internal memory size to reduce power consumption, as there is a trade-off between data transfer and memory size, and existing methods fail to efficiently manage this for given silicon area constraints.
Innovation Solution
A method that performs pre-analysis on a decimated image version to predict and transfer a smaller reference area, using techniques like 4:1 decimation, motion vector prediction, and selective data transfer based on search areas and memory constraints to minimize data transfer and enhance coding efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of energy
If internal memory size is increased to reduce data transfer, then power consumption is reduced, but silicon area increases
Solution Approach 1:
The patent performs pre-analysis on a decimated version of the image before the actual motion estimation process. This preliminary analysis identifies the reference area that needs to be transferred, allowing the system to minimize data transfer by only transferring necessary regions. The pre-analysis includes determining motion vectors and identifying reference areas based on a smaller decimated image, which then guides the data transfer process.
Solution Approach 2:
The patent extracts only the necessary reference area from the full image for transfer to internal memory. Instead of transferring the entire image or large portions of it, the system identifies and extracts specifically the reference area that is needed for motion estimation, based on the pre-analysis results. This extraction principle reduces the amount of data transferred while maintaining coding efficiency.
2Loss of energy
If data transfer is minimized by increasing internal memory size, then power consumption decreases, but device complexity increases
Solution Approach 1:
The patent performs pre-analysis on a decimated version of the image before the actual motion estimation process. This preliminary analysis identifies the reference area that needs to be transferred, allowing the system to minimize data transfer by only transferring necessary regions. The pre_analysis includes determining motion vectors and identifying reference areas based on a smaller decimated image, which then guides the data transfer process.
Solution Approach 2:
The patent extracts only the necessary reference area from the full image for transfer to internal memory. Instead of transferring the entire image or large portions of it, the system identifies and extracts specifically the reference area that is needed for motion estimation, based on the pre_analysis results. This extraction principle reduces the amount of data transferred while maintaining coding efficiency.
3Productivity
If pre_analysis is performed on full-resolution image, then coding efficiency improves, but data transfer and power consumption increase
Solution Approach 1:
The patent segments the image processing into two stages: first processing a decimated (down-sampled) version of the image to obtain motion information, then using this information to guide processing of the full-resolution image. This segmentation allows the computationally intensive pre_analysis to be performed on a smaller image, reducing data transfer and power consumption while still enabling accurate motion estimation for the full-resolution video.
Solution Approach 2:
The patent changes the resolution dimension by using a decimated version of the image for pre_analysis. Instead of working directly with the full-resolution image during the preliminary analysis phase, the system transforms the problem to a lower resolution domain, performs the analysis, then applies the results to the original high-resolution data. This dimensional transformation significantly reduces the amount of data that needs to be transferred and processed in the pre_analysis stage.
Data Source
AI summary
A method and apparatus for reduction of reference data transfer and coding efficiency improvement. The method includes performing pre-analysis on a decimated version of an image, and utilizing the predictions of the pre-analysis to transfer smaller reference area.