Research on Robustness of Color Device Characteristic Methods Based on Artificial Intelligence

Abstract

The nature of device color characteristic methods is the mutual conversion of device-dependent color space and device-independent color space. This paper does the comparative study on the robustness of some color space conversion methods which are based on fuzzy control, dynamic subspace divided BP neural network identification method, and fuzzy and neural identification method, by defining the robustness of color space conversion model and evaluation method. The result shows that the device color characteristic methods which are based on fuzzy and neural identification method can make the feature of BP neural network combine with fuzzy control to greatly improve the robustness of model.