基于模糊控制理论和BP神经网络系统的颜色空间转换模型研究

数字印前系统中色彩管理是一个重要的环节,而颜色空间的转换则是色彩管理的核心,特别是由设备相关颜色空间RGB到设备无关颜色空间CIELAB的转换。将模糊控制理论引入颜色空间转换模型,取得了较好的转换效果但仍存在色差大等缺点。本研究提出采用模糊控制模型将输入颜色空间动态划分为若干子空间,在子空间内部利用BP神经网络对颜色空间转换模型进行优化,得到了基于模糊神经系统的颜色空间转换模型。实验表明,该改进模型相对于单纯的基于模糊控制的转换模型而言,转换精度有很大提高。

Study on Color Space Transform Model Based on Fuzzy Control Theory and BP Neural Network

Color management is an important link in digital prepress system, and color space transfrom is the core of color management, especially from the device - related color space RGB to device - independent color space CIELAB. In this study, the improved color space transform model was introduced based on the fuzzy control and BP neural network system. The input color space was divided into many subspaces based on fuzzy control theory, and in the subspaces, the BP neural network was used to optimize the color space transform model. The analytical results showed that the converting accuracy of the proposed model is higher than the either color space transform model based on the fuzzy control theory.