Video Denoising via Noise Correlation Feedback
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional temporal denoising systems for videos suffer from 'speckle' noise issues when noise characteristics of adjacent pixels are correlated, leading to inconsistent filtering weights and poor image quality, especially with video sources that have been pre-processed like de-interlaced or scaled videos.
Innovation Solution
A video denoising system that estimates noise correlation between adjacent pixels using correlation coefficients for horizontally, vertically, and diagonally adjacent pixels, adjusting the maximum filtering weight to control the range of filtering weights and reduce 'speckle' noise occurrence.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional temporal denoising system uses threshold comparison method for motion detection, then motion detection can be performed, but speckle noise occurs due to inconsistent filtering weights in still regions
Solution Approach 1:
The patent introduces a feedback mechanism where the filtering weight calculation incorporates both motion probability and noise correlation information. The system continuously adjusts filtering weights based on the calculated noise correlation coefficients, creating a closed-loop control that adapts to the actual noise characteristics of the video signal, thereby eliminating speckle noise while maintaining denoising consistency.
Solution Approach 2:
The patent changes the parameters used for filtering weight calculation by introducing noise correlation coefficients (rh, rv, rd) that quantify the correlation between adjacent pixels. This parameter change allows the system to adaptively adjust filtering weights based on actual noise characteristics rather than relying solely on motion probability, resolving the speckle noise issue.
2Adaptability or versatility
If video sources undergo pre-processing like de-interlacing or scaling, then video compatibility is improved, but noise correlation between adjacent pixels increases causing speckle noise
Solution Approach 1:
The patent performs preliminary estimation of noise correlation coefficients before the main denoising operation. By calculating rh, rv, and rd values in advance and using them to adjust the maximum filtering weight, the system prepares for the correlated noise condition created by pre-processing operations, preventing speckle noise from occurring in the first place.
3Reliability
If maximum filtering weight is increased to improve denoising effect, then noise removal is enhanced, but speckle noise becomes more severe due to larger filtering weight fluctuations
Solution Approach 1:
The patent makes the maximum filtering weight dynamic by adjusting it according to the calculated noise correlation coefficients. Instead of using a fixed maximum weight, the system adaptively modifies it based on the actual noise characteristics, allowing optimal denoising performance while preventing excessive filtering weight fluctuations that cause speckle noise.
Data Source
AI summary
The present invention discloses a video denoising system based on noise correlation, said system comprises a unit for estimating correlation of noises of adjacent pixels, which receives input of the inter-frame difference and motion probability, and estimates correlation of noises of adjacent pixels according to correlation of inter-frame differences between adjacent pixels in a still region, and outputs a noise correlation coefficient; a maximum filtering weight adjusting unit, which adaptively adjusts the maximum weight for temporal filtering according to the noise correlation coefficient and outputs the maximum weight for temporal filtering; the maximum weight for temporal filtering can control the range of fluctuation of the temporal filtering weight and the difference between the denoising effects for different pixels. The system of the present invention can solve the problem of “speckle” noise occurred when the video noises have adjacent correlation in the conventional temporal denoising system for videos.