Low Light Noise Reduction via Pre-Demosaicing Patch Weighting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing noise suppression algorithms often generate repeating blotch patterns in low light areas of digital images, which are not effectively addressed by common noise reduction methods.
Innovation Solution
A low light noise reduction mechanism that performs denoising prior to demosaicing, using a weighted average based on similarity between image patches to determine appropriate values for both denoising and demosaicing, thereby minimizing noise patterns in low light areas.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If common noise suppression algorithms are applied to low light images, then noise is reduced, but repeating blotch patterns are generated
Solution Approach 1:
The patent performs denoising operation before demosaicing, using the denoised image to guide the subsequent demosaicing process. This preliminary denoising step prevents noise from being amplified during demosaicing, thereby avoiding blotch patterns while maintaining noise reduction effectiveness
Solution Approach 2:
The patent combines the denoising and demosaicing operations into a unified process by using the same set of similar patches and weighted averages for both operations. This integration ensures consistency between noise suppression and color reconstruction, preventing the generation of repeating blotch patterns
2Object-affected harmful factors
If denoising is performed before demosaicing using similar patches, then noise is reduced without blotch patterns, but computational complexity increases
Solution Approach 1:
The patent uses the same set of similar patches and weighted averages for both denoising and demosaicing operations. This multi-functional approach means that once the similar patches are identified, they serve dual purposes: suppressing noise and guiding color reconstruction, thereby reducing overall computational complexity compared to performing separate denoising and demosaicing operations
Solution Approach 2:
The denoising operation is performed first to produce a denoised image, which then serves as input for demosaicing. This preliminary processing simplifies the subsequent demosaicing operation by providing a cleaner input signal, reducing the computational burden of noise handling during the more complex demosaicing phase
Data Source
AI summary
A low light noise reduction mechanism may perform denoising prior to demosaicing, and may also use parameters determined during the denoising operation for performing demosaicing. The denoising operation may attempt to find several patches of an image that are similar to a first patch, and use a weighted average based on similarity to determine an appropriate value for denoising a raw digital image. The same weighted average and similar patches may be used for demosaicing the same image after the denoising operation.