Image Processing Apparatus Edge Artifact Prevention
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The Retinex method for dynamic range compression in images often results in overshoot or undershoot at edge portions due to sharp brightness changes, degrading image quality.
Innovation Solution
An image processing method that selects surrounding pixels based on a low-pass filter and pixel value differences to create a smoothed image, preventing overshoot and undershoot by excluding pixels with significant value differences from the reference pool, and combining this with the input image to produce an output image.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If the Retinex method is applied to compress dynamic range, then low-frequency areas achieve excellent compression results, but overshoot or undershoot occurs in high-frequency edge areas
Solution Approach 1:
The patent applies different processing strategies to different regions of the image. It calculates a first smoothed image using surrounding pixels for low-frequency compression, then identifies edge regions where brightness changes sharply. For these edge regions, it calculates a second smoothed image using only central pixels, avoiding the inclusion of surrounding pixels that would cause overshoot/undershoot. This local differentiation allows the method to maintain good compression quality in low-frequency areas while preventing edge artifacts.
Solution Approach 2:
The patent segments the image processing into distinct stages: first calculating a smoothed image using surrounding pixels, then identifying edge regions based on brightness changes, and finally processing different regions differently. The edge region is defined as a region where the brightness change exceeds a threshold, and this segmentation allows selective application of different smoothing methods to different parts of the image.
2Manufacturing precision
If surrounding pixels are used to calculate smoothed pixel values, then dynamic range compression is achieved, but edge regions experience sharp brightness changes causing quality degradation
Solution Approach 1:
The patent determines whether each region is an edge region by checking if brightness changes sharply. For non-edge regions, it uses surrounding pixels to calculate smoothed values for effective dynamic range compression. For edge regions, it uses only central pixels to avoid introducing artificial brightness changes. This local quality differentiation ensures reliable image quality across all regions while maintaining compression effectiveness.
Data Source
AI summary
An image processing apparatus includes a surrounding-pixels selecting unit that selects surrounding pixels that surround a correction pixel in an input image based on a low-pass filter, a reference-pixels selecting unit that selects pixels from the surrounding pixels based on pixel values of the correction pixel and the surrounding pixels, a smoothed-image creating unit that applies the low-pass filter to the reference pixels to create a smoothed image of the reference pixels, and an output image creating unit that creates an output image based on the smoothed image, and a difference image between the smoothed image and the input image.