Spatial-Temporal Noise Reduction and Contrast Enhancement in Digital Images
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing technologies struggle to effectively reduce noise and enhance contrast in digital images, particularly in the context of camera systems, where noise reduction methods are limited and contrast enhancement techniques are inadequate.
Innovation Solution
A method and system for spatial-temporal noise reduction and contrast enhancement that involves calculating spatial and temporal weights for each pixel, applying selective filtering based on these weights, and utilizing edge detection algorithms to enhance contrast, incorporating a three-dimensional noise reduction block and contrast enhancement block in the image processing pipeline.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If conventional noise reduction methods are applied, then noise is reduced to some extent, but image contrast deteriorates and edges become blurred
Solution Approach 1:
The patent applies different processing strengths to different regions of the image based on local characteristics. Edge regions receive minimal filtering to preserve sharpness, while uniform regions receive stronger noise reduction. This is achieved by detecting edge regions and adjusting the filtering strength accordingly, allowing noise reduction without compromising edge quality.
Solution Approach 2:
The filtering strength is dynamically adjusted based on local image characteristics rather than applying a uniform filter strength across the entire image. The system adapts the noise reduction intensity to match local variations in edge density and texture, enabling effective noise reduction in smooth areas while preserving edges in complex regions.
2Object-affected harmful factors
If stronger filtering is applied to reduce noise, then noise reduction improves, but image detail and contrast are lost
Solution Approach 1:
Different filtering strengths are applied to different regions based on their local characteristics. Regions with high detail content receive weaker filtering to preserve information, while uniform regions receive stronger filtering for effective noise reduction. This spatially adaptive approach prevents information loss in critical areas.
Solution Approach 2:
The patent applies filtering selectively rather than uniformly across the entire image. By identifying and protecting edge regions and high-detail areas from strong filtering while applying noise reduction where safe, the system achieves effective noise reduction without excessive information loss in critical image regions.
3Manufacturing precision
If conventional contrast enhancement techniques are used, then contrast is improved, but noise is amplified
Solution Approach 1:
The patent performs noise reduction before contrast enhancement in a sequential processing pipeline. By removing noise first and then applying contrast enhancement to the cleaned image, the system avoids amplifying noise during the contrast enhancement stage, achieving both good contrast and low noise levels.
4Object-affected harmful factors
If spatial filtering is applied to reduce noise, then noise is reduced, but temporal information and motion details are degraded
Solution Approach 1:
The patent separates spatial and temporal processing into distinct stages. Spatial noise reduction is applied to individual frames, and temporal filtering is applied separately to preserve motion information. This segmentation allows each processing stage to optimize for its specific function without interfering with the other, maintaining both noise reduction effectiveness and temporal fidelity.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Embodiments of the present invention disclose a system and method for reducing noise in a digital image. The method comprises calculating a spatial weight for each pixel s in a current frame, wherein said spatial weight is computed based on a summation of diffused values associated with the pixel s; and selectively filtering an intensity I of the pixel s spatially using said spatial weight.