Color Correction Matrix Tuning for Perceptual Hue Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current color correction matrix optimization methods fail to accurately match the colors in corrected images to human perception due to reliance on Euclidean distance and the tendency to converge to local minimums, leading to suboptimal color correction.
Innovation Solution
A method that determines a loss function based on human eye perception of hue and saturation, using Lab coordinates to construct a global hue error and saturation constraint regularization term, iteratively updating the color correction matrix to optimize it for better alignment with human perception.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If the Euclidean distance is used to optimize the color correction matrix, then the optimization process is simple and fast, but the color correction result does not match human eye perception
Solution Approach 1:
The patent changes the optimization parameter from Euclidean distance to a loss function based on human eye perception characteristics. The loss function incorporates hue error and saturation error terms that model how human eyes perceive color differences, allowing the optimization to converge to a solution that matches human perception rather than just minimizing mathematical distance.
Solution Approach 2:
The patent replaces the mechanical/mathematical Euclidean distance metric with a perceptual model that simulates human eye response. By substituting the standard mathematical distance measure with a perception-based loss function, the system achieves optimization that aligns with human visual perception rather than purely mathematical correctness.
2Ease of manufacture
If the least squares method is used to initialize the color correction matrix, then the initialization process is simple and efficient, but the optimization result falls into local minimum traps
Solution Approach 1:
The patent applies preliminary action by using the least squares method to obtain an initial color correction matrix before performing iterative optimization. This initial matrix serves as a starting point that is easy to compute, and then the iterative process refines it further. The key insight is that multiple initial matrices are generated and then optimized together, allowing the system to escape local minimums while maintaining the simplicity of least squares initialization.
Solution Approach 2:
The patent introduces dynamics by transitioning from a static initialization method to a dynamic iterative optimization process. The color correction matrix is updated repeatedly based on the loss function gradients, allowing the system to adapt and escape local minimums. The dynamic nature of iterative optimization enables the system to find global optimum solutions that static initialization methods cannot achieve.
3Device complexity
If traditional color correction methods are used, then the processing flow is simple, but severe color cast remains in the corrected image
Solution Approach 1:
The patent segments the color correction process into distinct components: white balance correction for color temperature, and color correction matrix optimization for color cast. By separating these functions and optimizing them independently with appropriate loss functions, the system achieves more accurate color correction without excessive complexity. The segmentation allows each component to be tuned specifically for its purpose.
Solution Approach 2:
The patent implements feedback through the iterative optimization process where the loss function continuously evaluates the color correction results and feeds this information back to adjust the color correction matrix. This feedback mechanism allows the system to progressively improve color correction accuracy by learning from the differences between predicted and actual human perception, eliminating severe color cast that would remain in one-shot correction methods.
Data Source
AI summary
Provided are a color correction matrix optimization method, an electronic device, and a medium. The method includes performing correction processing on a color source matrix of an acquired color chart image based on a color correction matrix to obtain an output color matrix; determining a global hue error and a saturation constraint regularization term based on Lab coordinates of the output color matrix and Lab coordinates of a target color matrix of a target color chart image and constructing a loss function; and updating the color correction matrix based on a value of the loss function until the value of the loss function converges to obtain an optimal color correction matrix to perform color correction on a to-be-processed image based on the optimal color correction matrix.