Image Compression Circuit Using Pixel Group Correlation Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image compression techniques, such as BTC and DCT, suffer from significant compression errors for specific types of images, leading to image deterioration, particularly in portable devices that display mixed content like characters and photographs, and increase circuit size due to complex correlation calculations.
Innovation Solution
An image processing circuit and method that generates representative values for groups of pixels, calculates correlations between these datasets, and selects an appropriate compression process to reduce compression errors while minimizing circuit size, using a pre-process to group pixels and perform bit-reducing processes on datasets to generate compressed image data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If correlation is calculated for every possible combination of two pixels to select compression method, then compression error is reduced, but circuit size increases
Solution Approach 1:
The patent segments the pixel set into multiple groups, where each group has a predetermined number of pixels. Instead of calculating correlations for all possible combinations of individual pixels, the system calculates correlations between groups. This segmentation reduces the number of correlation calculations from O(N²) for N pixels to O((N/k)²) for N/k groups, significantly reducing circuit size while maintaining compression quality.
Solution Approach 2:
The patent performs correlation calculations only for groups of pixels rather than all individual pixel combinations. This partial action approach calculates correlations for a subset of pixel relationships (group-level correlations) rather than exhaustive pixel-level correlations, achieving sufficient compression error reduction with reduced computational complexity and smaller circuit size.
2Quantity of substance
If block coding is used to compress image data, then data transmission amount is reduced, but image deterioration occurs for specific image types
Solution Approach 1:
The patent implements dynamic selection of compression methods based on correlation calculations. Instead of using a fixed block coding method (BTC or DCT) for all images, the system dynamically determines the appropriate compression method by calculating correlations between pixel groups and selecting the method that minimizes compression error for that specific image content. This dynamic adaptation prevents image deterioration for both low-frequency (photographs) and high-frequency (characters) images.
Solution Approach 2:
The patent changes the compression parameter (compression method) based on image characteristics. By calculating correlations and using them to select between different compression methods (BTC for low correlation, DCT for high correlation), the system adapts the compression parameters to match the image content, thereby maintaining image quality while achieving data compression.
Data Source
AI summary
An image processing circuit includes: a representative-values calculation circuit and an all-combinations comparing compression circuit. The representative-values calculation circuit is configured to generate M datasets each including a plurality of representative values by performing a pre-process on image data associated with said N pixels, M being a natural number more than one and less than N. The all-combinations comparing compression circuit is configured to calculate correlations between two datasets selected from said M datasets for all possible combinations of the two datasets, to select a compression process from a plurality of compression processes in response to the calculated correlations, and to generate said compressed imaged data by compressing said M datasets by using said selected compression process. The image processing circuit may be incorporated in a display panel driver.