Image Color Correction Using Confidence-Weighted Region Values
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing color correction methods for digital images often result in unnatural color balances due to focusing solely on the face area, which can vary significantly between individuals.
Innovation Solution
An image processing apparatus and method that calculates multiple color correction values using different processing methods, assigns confidence levels to these values based on their reliability, and combines them using significance levels to produce a new color correction value for more accurate and user-preferred color balancing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If color correction is performed by focusing only on the face area (skin color area), then the color correction can be performed on a specific region of interest, but the color correction becomes unnatural because the face area differs in individuals
Solution Approach 1:
The patent segments the image into multiple regions including face area, sky area, and grass area, and performs color correction for each region separately using region-specific correction values. This allows the system to focus on the face area when present while also considering other regions, thereby maintaining naturalness across the entire image while still providing targeted correction where needed.
2Reliability
If multiple color correction values are calculated by different processing methods, then the reliability of color correction can be improved, but the device complexity increases
Solution Approach 1:
The patent combines multiple color correction values calculated from different regions (face area correction value, sky area correction value, grass area correction value) into a single comprehensive correction approach. Each region's correction value is calculated independently using appropriate methods, then merged to produce the final color correction, thereby improving reliability without requiring a completely separate system for each method.
Solution Approach 2:
The patent changes the parameters used for color correction based on the detected image content. Different correction values are calculated using different parameters depending on the region being analyzed (e.g., skin tone parameters for face area, color temperature parameters for sky area). This allows the system to adapt its processing approach dynamically, improving reliability while managing complexity through parameter adaptation rather than structural complexity.
Data Source
AI summary
Fed image data is used, to respectively calculate two color correction values in two correction value calculating circuits. Further, a target skin color chromaticity value and significance levels indicating which of the two color correction values should be attached importance to are inputted from an input device. Respective confidence levels for the two color correction values are calculated on the basis of the two color correction values and the inputted target skin color chromaticity value. On the basis of the calculated confidence levels, the inputted significance levels, and the two color correction values, both of the confidence levels and the significance levels are reflected, to calculate a new color correction value in a color correction value calculating circuit.