Image Enhancement via Pixel Headroom Distribution
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image enhancement techniques, such as the Burt-pyramid algorithm, often result in overshoots and require clipping to prevent image distortion, leading to poor quality, especially in areas with edges.
Innovation Solution
A multi-scale approach is used to modify pixel values in an input image by calculating a pixel headroom signal and distributing it across sub-images, allowing for controlled enhancement without significant overshoot or clipping, while prioritizing sub-image processing based on available headroom and frequency bands.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If pixel values are enhanced uniformly across all frequency ranges, then sharpness and contrast are improved, but overshoot occurs at edges requiring clipping that degrades image quality
Solution Approach 1:
The image is divided into multiple frequency bands (sub-images) through wavelet transformation, allowing differential enhancement strategies for different frequency ranges. High-frequency components are enhanced more aggressively for sharpness while low-frequency components are enhanced more conservatively to avoid overshoot at edges.
Solution Approach 2:
The enhancement factor is made spatially variable rather than uniform. The patent calculates local statistics (mean and standard deviation) for different regions and applies adaptive enhancement factors that depend on local image characteristics, preventing overshoot in edge regions while maintaining enhancement in smooth areas.
2Manufacturing precision
If enhancement factors are increased to maximize contrast improvement, then image contrast is enhanced, but the amount of clipping increases leading to distortion
Solution Approach 1:
The patent performs a preliminary analysis of the image to calculate local statistics (mean and standard deviation) before applying enhancement. This preliminary characterization of image regions allows the enhancement factor to be pre-adjusted for each region, preventing overshoot and clipping before they occur.
Solution Approach 2:
The enhancement factor is changed from a uniform constant to a spatially varying parameter that depends on local image statistics. The patent dynamically adjusts the enhancement factor for each pixel or region based on local mean and standard deviation, allowing higher enhancement in safe regions and lower enhancement near edges.
Data Source
Figure 1
Figure 2
Figure 3~4
AI summary
The invention relates to an apparatus and a method for computing an output image (145) on the basis of an input image (100). The method comprises the steps of splitting (105) the input image (100) into multiple sub-images, each sub-image comprising information from a respective frequency band of the input image (100), calculating (110) a pixel headroom signal quantifying the margin available for enhancement of pixel values of pixels in the input image (100), calculating (115) pixel enhancement factors for pixels within sub-images to be modified in dependence upon the pixel headroom signal, such that the margin as quantified by the pixel headroom signal is spread across the respective sub-images to be modified, modifying (120) the respective sub-images by using the pixel enhancement factors and the corresponding pixel values from the respective sub-images, and generating (135) the output image (145) by using the modified sub-images. The present invention also relates to an image signal comprising pixel values forming an image (100) and a set of pixel enhancement factors allowing enhancement of the image.