Image Processing Apparatus Noise Reduction Edge Preservation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing noise reduction techniques, such as bilateral and ε filters, complicate processing and increase operation costs, especially in real-time applications, and often result in non-continuity of image quality by smoothing small-amplitude textures and edges.
Innovation Solution
An image processing apparatus calculates the absolute difference in pixel values within a pixel block, performs weighted averaging based on these differences, and applies smoothing while maintaining edge components and small-amplitude textures using a simplified filter operation that reduces operation costs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If bilateral filter is used for noise reduction, then random noise is reduced while maintaining edges, but processing becomes complicated and operation cost increases
Solution Approach 1:
The patent segments the complex bilateral filter operation into two simpler stages: first applying a standard spatial filter for noise reduction, then selectively correcting pixels near edges by comparing with neighboring pixels. This segmentation reduces the computational complexity while maintaining edge preservation capability.
Solution Approach 2:
The patent introduces an intermediary step of detecting edge pixels through gradient calculation, which acts as a mediator between the simple spatial filter and the final output. This intermediary mechanism allows the system to apply different processing to different regions without the full complexity of a bilateral filter.
2Manufacturing precision
If c filter is used for noise reduction, then random noise is reduced while maintaining edges, but small-amplitude texture is smoothed and non-continuity occurs in image quality
Solution Approach 1:
The patent applies local quality by using gradient magnitude to identify edge pixels and applying correction only to those specific locations. This localized approach preserves small-amplitude textures in non-edge regions while maintaining edges, avoiding the blanket smoothing effect of the c filter.
Solution Approach 2:
The patent dynamically changes the processing parameter (correction application) based on the gradient magnitude at each pixel location. Pixels with high gradient magnitudes (edges) receive correction, while pixels with low gradient magnitudes (textures) are left unchanged, thus preserving texture information.
3Manufacturing precision
If standard spatial filter is used for noise reduction, then random noise is reduced, but image edges blur and details are lost
Solution Approach 1:
The patent implements feedback by calculating the gradient magnitude after the spatial filter operation and using this information to identify and correct edge pixels. This feedback mechanism allows the system to detect and preserve edges that would otherwise be blurred by the spatial filter.
Solution Approach 2:
The patent performs preliminary noise reduction using the spatial filter, then subsequently corrects edge pixels based on gradient analysis. This two-stage approach with preliminary action allows noise reduction to be applied first, followed by selective edge restoration.
Data Source
AI summary
An image processing apparatus includes a memory device and at least one processor to calculate, in a pixel block including a pixel of interest and a plurality of surrounding pixels, an absolute value of a difference in pixel value between the pixel of interest and a surrounding pixel, the calculating being performed for each of the surrounding pixels, to convert a pixel value of the surrounding pixel in accordance with the calculated absolute value, the converting being performed for each of the plurality of surrounding pixels, and to perform smoothing processing for the pixel of interest by using the plurality of surrounding pixels including the surrounding pixel whose pixel value is converted. The converting performs a combination of a pixel value of the pixel of interest and the pixel value of the surrounding pixel, so that as the calculated absolute value increases, a weight of the pixel of interest increases.