Correlation Value Calculation Device Circuit Scale Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing correlation value calculation devices for image processing, used in noise reduction and motion vector detection, face challenges in reducing circuit scale while maintaining high-speed correlation value calculations, particularly as the degree of parallelism and block size increase.
Innovation Solution
A correlation value calculation device that selects a pixel of interest and calculates correlation values between base and reference blocks of varying sizes, using intermediate calculation units to compute values for specific block sizes and interpolating for other sizes, thereby reducing circuit scale and enhancing processing efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If correlation value calculation units are provided in parallel to increase calculation speed, then the correlation value calculation speed is improved, but the circuit scale increases
Solution Approach 1:
The patent segments the correlation value calculation into two distinct phases: a calculation phase where correlation values are computed for minimum block sizes using multiple parallel calculation units, and an interpolation phase where correlation values for other block sizes are derived through interpolation. This segmentation allows the system to maintain high calculation speed through parallel processing while reducing overall circuit scale by avoiding the need for separate calculation units for all block sizes.
Solution Approach 2:
The patent changes the block size parameter dynamically during the calculation process. Instead of having fixed calculation units for each block size, the system calculates correlation values for minimum block sizes and then interpolates to obtain values for larger block sizes. This parameter change approach enables the system to achieve high calculation speed for minimum block sizes while reducing circuit scale by reusing the same calculation units for different block sizes through interpolation.
2Productivity
If the degree of parallelism and block size are increased to improve calculation speed, then the correlation value calculation speed is improved, but the circuit scale increases
Solution Approach 1:
The patent applies partial action by calculating correlation values only for minimum block sizes using parallel calculation units, and then using interpolation to derive values for larger block sizes. This approach uses calculation resources efficiently by performing calculations only where necessary (for minimum block sizes) and avoiding redundant calculation units for larger block sizes that can be obtained through interpolation.
Solution Approach 2:
The patent introduces interpolation as an intermediary mechanism between the parallel calculation units and the final correlation values for various block sizes. The interpolation unit acts as a mediator that transforms the calculated minimum block size correlation values into correlation values for other block sizes, thereby eliminating the need for separate calculation units for each block size while maintaining high calculation speed.
Data Source
AI summary
A correlation value calculation device for selecting a pixel of interest within an image and a plurality of reference pixels within the image or a different image and calculating correlation values between a base block including the pixel of interest at a prescribed position and a plurality of reference blocks including each reference pixel at the prescribed position and having the same block size as the base block includes a block size determination unit, first intermediate correlation value calculation units configured to calculate correlation values in a plurality of types of block sizes, second intermediate correlation value calculation units configured to calculate correlation values in some block sizes among the plurality of types of block sizes, and a correlation value interpolation unit configured to obtain correlation values according to an interpolation.