Pixel-by-Pixel Visual Processing for Noise Suppression in Wide Dynamic Range Images
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing devices face challenges in performing visual processing on images with wide dynamic ranges, as they often enhance noise levels in bright areas and fail to adequately suppress noise in pixels with varying luminance values, leading to image deterioration.
Innovation Solution
An image processing device that performs visual processing on a pixel-by-pixel basis using human visual characteristics, incorporating a spatial processing section, visual processing section, and input signal processing section to restore resolution and correct signal deterioration, with adaptive noise reduction based on gain values determined by a two-dimensional LUT.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Illumination intensity
If visual processing is performed on image data with wide dynamic range to increase brightness in dark portions, then the brightness of dark areas is improved, but the noise component is also increased
Solution Approach 1:
The patent applies local quality by performing visual processing on a pixel-by-pixel basis rather than uniformly across the entire image. The gain for each pixel is determined individually based on its luminance value and the luminance distribution of surrounding pixels, allowing brightening of dark areas while avoiding excessive noise amplification in specific regions.
Solution Approach 2:
The patent changes the gain parameter adaptively for each pixel based on its luminance characteristics. By using a two-dimensional LUT that maps combinations of pixel luminance and surrounding luminance distribution to appropriate gain values, the system dynamically adjusts the processing strength to balance brightness enhancement and noise control.
2Device complexity
If noise reduction processing is performed with uniform strength on all pixels in a block, then processing is simplified, but edges of areas with small noise amplitude become dull
Solution Approach 1:
The patent applies local quality by determining noise reduction strength on a pixel-by-pixel basis rather than uniformly across blocks. Each pixel's noise reduction gain is calculated based on its specific luminance characteristics and surrounding distribution, preserving edge sharpness in areas with small noise amplitude while effectively reducing noise in areas with high noise content.
3Measurement precision
If block size is reduced to improve noise reduction precision, then noise suppression accuracy is improved, but the amount of data to be processed increases
Solution Approach 1:
The patent changes the processing approach by using pixel-by-pixel visual processing with gain determination based on luminance characteristics rather than block-based processing. This method achieves precise noise suppression adapted to each pixel's characteristics without the computational overhead of processing many small blocks, as the processing efficiency is improved through efficient gain calculation using luminance distribution analysis.
Data Source
AI summary
A visual processing section 120 is provided which performs visual processing on a pixel-by-pixel basis according to a predetermined function defining a relation between a pixel value of a target pixel which is a pixel under visual processing and a pixel value of the target pixel after subjected to the visual processing. An input signal processing section 150 is further provided which performs, on a pixel-by-pixel basis, at least one of signal processing to restore resolution of an image data input to the visual processing section 120 and signal processing to correct signal deterioration due to the visual processing. In the input signal processing section 150, the at least one signal processing is performed according to a gain determined by using the function.