Image Processing Apparatus Contrast Expansion via Sensitivity Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing methods fail to adequately enhance local contrast between low-sensitivity and high-sensitivity images, resulting in poor visibility and loss of information, especially in dark and bright areas.
Innovation Solution
An image processing apparatus that generates contrast-expanded images by applying smoothing filters, dividing pixel values, and combining brightness and contrast components from low-sensitivity and high-sensitivity images using specific weight coefficients and multipliers to emphasize local contrast.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Illumination intensity
If a low-sensitivity image is used to pick up bright objects, then the contrast in bright areas is high, but the dark areas become blackened with large blackened areas and much noise
Solution Approach 1:
The image is segmented into multiple sensitivity regions, with some pixels capturing low-sensitivity images (for bright areas) and other pixels capturing high-sensitivity images (for dark areas), allowing each region to be optimized for its specific luminance range
Solution Approach 2:
The patent combines multiple images with different sensitivities into a single composite image, merging the advantages of low-sensitivity images (high contrast in bright areas) and high-sensitivity images (preserved information in dark areas) to produce an image with uniformly high quality across all luminance ranges
2Illumination intensity
If a high-sensitivity image is used to pick up dark objects, then the contrast in dark areas is high, but the bright areas become saturated in white
Solution Approach 1:
The image sensor is divided into regions with different sensitivities, where certain pixels are optimized for capturing dark areas (high sensitivity) while other pixels are optimized for capturing bright areas (low sensitivity), preventing saturation while maintaining contrast where needed
Solution Approach 2:
The patent merges high-sensitivity and low-sensitivity images to create a composite image that preserves detail in both dark and bright areas, eliminating the saturation problem while maintaining the high contrast benefits in dark regions
3Loss of information
If the high-frequency component of the low-sensitivity image is emphasized, then noise in dark areas is enhanced, but if it is not emphasized, then local contrast in dark areas cannot be improved
Solution Approach 1:
The processing is segmented by sensitivity type, with low-sensitivity images processed to preserve local contrast in bright areas and high-sensitivity images processed to preserve local contrast in dark areas, with selective frequency emphasis applied to each based on its strength
Solution Approach 2:
Different processing strategies are applied to different regions: low-sensitivity images undergo high-frequency emphasis for bright areas where contrast is needed, while high-sensitivity images maintain their natural characteristics in dark areas to avoid noise amplification, with each region processed according to its specific requirements
4Illumination intensity
If the contrast is calculated using pixel values in the periphery of a target pixel, then local contrast is emphasized, but in blackened areas or white saturated areas where pixel values are uniform, the contrast becomes 1 and cannot be enhanced
Solution Approach 1:
The patent segments the image processing by sensitivity type, calculating contrast separately for low-sensitivity and high-sensitivity images, and selectively applying contrast enhancement only where it is effective, avoiding the problem of uniform areas where contrast calculation fails
Solution Approach 2:
Instead of calculating contrast from peripheral pixels (which fails in uniform areas), the patent inverts the approach by using the original pixel values directly from the appropriately selected sensitivity image, preserving the actual luminance information rather than deriving it from potentially misleading contrast calculations
Data Source
AI summary
An image processing apparatus generates a brightness component image of a low-sensitivity image by applying a smoothing filter to the low-sensitivity image, generates a brightness component image of a high-sensitivity image by applying the smoothing filter to the high-sensitivity image, generates a contrast component image of the low-sensitivity image by dividing the low-sensitivity image by the brightness component image of the low-sensitivity image, generating a contrast component image of the high-sensitivity image by dividing the high-sensitivity image by the brightness component image of the high-sensitivity image, generates a combined brightness component image by combining the brightness component images of the low-sensitivity image and of the high-sensitivity image, generates a combined contrast component image by combining the contrast component images of the low-sensitivity image and of the high-sensitivity image, and finally generates a contrast-expanded image by multiplying the combined brightness component image and the combined contrast component image.