Digital Image Noise Reduction Using Adaptive Edge Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing digital image noise reduction methods often blur and damage image edges, failing to effectively separate noise from image content while maintaining image quality.
Innovation Solution
A method and apparatus that utilize high pass filtering to identify edge regions, low pass filtering on non-edge regions with specific luminance criteria, and sigma filtering, adjusting mask sizes based on luminance deviations to isolate and reduce noise without harming image edges.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If conventional noise reduction methods are applied to digital images, then noise is reduced, but image edges are blurred and damaged
Solution Approach 1:
The image is segmented into edge regions and non-edge regions using high pass filtering. Different filtering operations are then applied to different regions: edge regions are preserved without aggressive noise reduction, while non-edge regions undergo noise reduction processing. This segmentation allows simultaneous noise reduction and edge preservation by treating different image regions differently.
Solution Approach 2:
The patent applies different filtering characteristics to different spatial locations in the image. Edge regions receive minimal filtering to preserve their sharpness and integrity, while non-edge regions receive stronger filtering to remove noise. This local differentiation of filtering quality enables the system to reduce noise overall while protecting edge structures from degradation.
2Object-affected harmful factors
If noise reduction filtering is applied uniformly across the entire image, then noise is reduced, but edge regions lose their sharpness and detail
Solution Approach 1:
The filtering operation is made dynamic and adaptive rather than static and uniform. The system dynamically determines which regions are edges and which are non-edges, then dynamically adjusts the filtering strength applied to each region. This dynamic approach allows the filtering to adapt to local image characteristics, preserving edge sharpness while reducing noise in appropriate regions.
Solution Approach 2:
The patent changes the filtering parameters (filtering strength, filter type) based on the detected image content. In edge regions, filtering parameters are adjusted to preserve sharpness and detail. In non-edge regions, parameters are adjusted to maximize noise reduction. This parameter adaptation enables the system to optimize both edge sharpness and noise reduction performance.
3Object-affected harmful factors
If aggressive noise reduction is applied to improve image quality, then noise is removed, but image contours and patterns are distorted
Solution Approach 1:
The image is segmented into edge regions and non-edge regions using high pass filtering. Different filtering operations are then applied to different regions: edge regions are preserved without aggressive noise reduction, while non-edge regions undergo noise reduction processing. This segmentation allows simultaneous noise reduction and edge preservation by treating different image regions differently.
Solution Approach 2:
The patent applies different filtering characteristics to different spatial locations in the image. Edge regions receive minimal filtering to preserve their sharpness and integrity, while non-edge regions receive stronger filtering to remove noise. This local differentiation of filtering quality enables the system to reduce noise overall while protecting edge structures from degradation.
Data Source
AI summary
There are provided a method and an apparatus for reducing noise in a digital image capable of reducing noise while preventing damage to an edge of a digital image. The apparatus includes: a high pass filtering unit determining an edge region of an input image; a low pass filtering unit performing low pass filtering on a region of the input image determined not to be the edge region by the high pass filtering unit; and a sigma filtering unit performing sigma filtering on the region of the input image determined not to be the edge region by the high pass filtering unit.