The Color Characteristic Model Based on Optimized BP Neural Network

Abstract

Due to BP neural network’s own limitations, it requires a large number of samples for complex prediction, and its generalization ability is weak. This paper puts forward an optimized BP neural network algorithm. The methods focus on data normalization in order to improve maximum speed limit, the inertia constant, and fitness function to eventually optimize BP neural network weight value and threshold value to reduce its distribution range before using BP neural network for color prediction. Reducing the possibility of BP neural network prediction model getting into local optimization has good nonlinear fitting capability and higher prediction accuracy of color space conversion. If using in the ICC profile, it will ensure the accuracy of the color conversion.