用模糊ART神经网设计自适应调色板

Adaptive Palette Designed by Fuzzy ART Neural Network

本文分别采用模糊ＡＲＴ神经网和模糊ＡＲＴ神经网的变体设计了自适应调色板．文中通过实验证明，模糊ＡＲＴ神经网设计灵活，且客观映射误差有比较明显地改善；同时，通过负模糊度概念的引入，构筑模糊ＡＲＴ神经网的一种变体：负模糊ＡＲＴ，进而从实验上证明了负模糊ＡＲＴ。

In this paper an adaptable palette is designed by fuzzy ART neural network and its modified fuzzy ART neural network is presented. It is experimentally shown that the designing fuzzy ART neural network is flexible and the mapping error is obviously improved. In addition, this paper proposes a new concept about negative fuzzy logic. The negative fuzzy logic neural network constructed by negative logic can classify less categories so as to obtain better result of vision.