摘要
设备颜色特性化是色彩管理技术得以顺利实现的关键技术之一,其核心是设备相关颜色空间和与设备无关颜色空间之间的相互转换。本研究首先定义颜色空间转换方法的鲁棒性概念及其评价方法,在此基础上,对基于模糊控制的颜色空间转换方法、基于动态子空间划分的BP神经网络颜色空间转换方法和基于模糊神经辨识的颜色空间转换方法等基于人工智能的颜色空间转换方法的鲁棒性做了比较研究。研究结果显示:基于模糊神经辨识的颜色空间转换方法能够结合BP神经网络和模糊控制的特点,使其鲁棒性得到很大的提高。
Device color characteristic is one of the key technologies of color management technology. The core of device col- or characteristic methods is the mutual conversion between device - dependent color space and device - independent color space. In this study, after mading a definition for the robustness of color space conversion method and evaluation method, the comparative study on the robustness of some color space conversion methods included conversion method based on fuzzy control, BP neural network identification method based on dynamic subspace dividing, and fuzzy neural identification meth- od. The results showed that the device color conversion method based on fuzzy neural identification combining with the char- acteristics of fuzzy control and BP neural network can greatly improve the robustness of method.
出处
《中国印刷与包装研究》
CAS
2012年第6期5-9,共5页
China Printing Materials Market
基金
陕西科技大学科研启动基金项目(博士启动基金)--人工智能设备颜色特征化方法的应用研究(No.BJ12-25)
关键词
人工智能
颜色空间转换
鲁棒性
稳定性
Artificial intelligence
Color space conversion
Robustness
Stability
