Relative Histograms for Independent Color Channel Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for color correction in digital images, such as histogram stretching, often fail to accurately adjust color channels independently, leading to unintended tone changes, particularly in skin tones, due to assumptions about similarity in statistical measures across channels.
Innovation Solution
The use of relative histograms, specifically one-dimensional and two-dimensional relative histograms, to compare and correct color channels by establishing relationships between primary color channels (R, G, and B) without assuming similar intensity distributions, allowing for targeted adjustments based on the ratios and ratios' reciprocals to maintain accurate color balance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional histogram stretching is applied to correct color channels independently using a common acceptable rejection percentage, then the overall tone of the image is improved, but the skin tones become unacceptably distorted
Solution Approach 1:
The patent segments the color correction process by creating separate correction pathways for different regions of the image. It identifies skin tone regions and applies different histogram stretching parameters to skin tones versus non-skin areas, allowing independent optimization of correction accuracy for each region without mutual interference
Solution Approach 2:
The patent implements local quality by applying spatially varying correction strategies. Different parts of the image (skin tones vs. non-skin areas) receive tailored histogram stretching parameters based on their local characteristics, ensuring that correction accuracy is optimized locally for each region's specific requirements
2Ease of operation
If the histograms are stretched using linear transformations with a single tolerable rate for all channels, then the processing simplicity is maintained, but the color balance accuracy deteriorates
Solution Approach 1:
The patent introduces dynamic adaptation by automatically detecting skin tone regions and adjusting histogram stretching parameters based on local image characteristics. The system transitions from static, uniform correction parameters to dynamic, region-specific parameters that adapt to the actual content being processed, improving color balance accuracy without requiring manual intervention
3Ease of manufacture
If a filter is applied to correct color in direct digital photography, then the color correction is applied uniformly across the image, but the ability to preserve local tone characteristics is lost
Solution Approach 1:
The patent segments the image into different regions (skin tones and non-skin areas) and applies different correction filters to each segment. This allows the system to maintain the ease of automated filter application while preserving local tone characteristics through region-specific correction parameters
Solution Approach 2:
The patent implements local quality by making the correction filter's parameters dependent on the local image content. Skin tone regions receive different correction parameters compared to non-skin regions, allowing the filter to preserve local tone characteristics while still providing automated correction
Data Source
AI summary
Relative histograms compare occurrences in event channels for a multi-channel data set for determining comparisons therebetween. One or more relative histograms are formed using relating functions applied to channels, preferably including a quotient between a first reference channel and a second channel and between the reference channel and at least a third channel. Preferably the relating functions generate values where a ratio and its reciprocal are symmetrical about an identity value. More preferably a data set is presented in a two-dimensional histogram for establishing the deviation of significant counts from an adjustment point which is preferably at the identity value. In practice, the relative histograms can be applied to correct tint in the red, green and blue channels of color images.