Compressed Image Denoising Through Quality Factor Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
High-resolution video compression results in significant reduction of image quality, leading to increased transmission and storage costs, and existing methods fail to effectively remove compression noise in compressed images.
Innovation Solution
An image processing method involving feature extraction, reconstruction, quality factor prediction, and denoising using neural networks to generate a denoised compressed image with reduced noise.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of energy
If video compression is applied to reduce transmission and storage costs, then transmission and storage costs are reduced, but image quality deteriorates significantly
Solution Approach 1:
The patent segments the image processing into multiple stages: compression, quality factor prediction, and selective denoising. By dividing the processing pipeline, the system can apply different operations to different regions based on compression artifacts, thereby maintaining overall image quality while preserving compression benefits.
Solution Approach 2:
The patent implements local quality enhancement by predicting quality factors for different regions of the compressed image and applying denoising operations selectively. Regions with higher compression artifacts receive more aggressive denoising, while regions with better quality preserve more original compressed data, achieving localized optimization of image quality.
2Productivity
If compression rate is increased to reduce data size, then transmission and storage efficiency improves, but compression noise increases
Solution Approach 1:
The patent employs feedback mechanisms by predicting quality factors from the compressed image itself and using these predictions to guide the denoising process. The system continuously monitors compression quality and adjusts denoising intensity accordingly, creating a closed-loop system that adapts to varying compression levels and noise characteristics.
Solution Approach 2:
The patent changes key parameters including quality factor thresholds, denoising strength, and processing intensity based on the predicted compression quality. By dynamically adjusting these parameters according to the actual compression level and artifact distribution, the system optimizes the balance between noise removal and detail preservation.
Data Source
AI summary
An image processing method of an electronic device includes performing feature extraction on a compressed image to obtain a compression feature map of the compressed image; performing reconstruction on the compression feature map to obtain a reconstruction feature map of the compressed image; performing quality factor (QF) prediction on the compression feature map to obtain a QF of the compressed image; and generating a denoised compressed image having compression noise reduced or removed from the compressed image by performing denoising on the reconstruction feature map based on the QF.