Frequency Component Compensation for Image Compression
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current lossy image compression techniques, such as JPEG and JPEG-2000, face challenges in achieving a higher compression rate without degrading image quality, as they primarily discard higher frequency components, which are less sensitive to human eyes, while retaining lower frequency components that occupy most of the data size.
Innovation Solution
The method involves estimating transform coefficients of a frequency component for an image, performing a dot multiplication operation with a basis function to generate a compensation image, and combining the original image with the compensation image to achieve a higher compression rate without significantly degrading image quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If higher frequency components are discarded to achieve compression, then compression rate increases, but image quality degrades
Solution Approach 1:
The image is divided into frequency components through transform (DCT/DST), separating the image into distinct frequency bands that can be independently processed. This segmentation allows selective compression of high-frequency components while preserving low-frequency components that contain most visual information, resolving the contradiction between compression rate and image quality.
Solution Approach 2:
Different quantization strength is applied to different frequency components based on human visual sensitivity. Low-frequency components retain higher precision while high-frequency components undergo stronger quantization. This local quality differentiation enables effective compression without significant image quality degradation.
2Quantity of substance
If lower frequency components are removed to increase compression, then data size reduces, but image quality significantly degrades
Solution Approach 1:
The patent dynamically adjusts quantization parameters based on image content characteristics and desired compression rate. By changing quantization parameters adaptively, the system can achieve higher compression ratios while maintaining acceptable image quality, avoiding the significant degradation that would occur with fixed aggressive quantization of low-frequency components.
Solution Approach 2:
The quantization table is dynamically generated and adjusted based on the specific image being compressed and the target compression rate. This dynamic adaptation allows the system to optimize the balance between data size reduction and image quality preservation for each individual case, rather than applying static quantization that would inevitably degrade quality.
3Manufacturing precision
If machine learning is used to improve image quality, then image quality increases, but computational complexity increases
Solution Approach 1:
The patent uses transform coefficients as an intermediary representation that captures image information in a compressed form. By performing quantization and compression operations in the transform domain rather than the spatial domain, the system achieves effective compression with reduced computational complexity compared to applying machine learning directly to full-resolution images.
Data Source
AI summary
The present disclosure relates to a method for compensating an image. The method comprises estimating transform coefficients of a frequency component for a first image based on the first image, performing a dot multiplication operation between the estimated transform coefficients and a basis function associated with the frequency component to generate a compensation image, and combining the first image and the compensation image to generate a combined image.