基于模糊神经网络的静态手语词汇识别(英文)  

A Recognition Method for Static Words of Chinese Sign Language Based on Fuzzy-Neuro Network

作　　者：邹伟[1] 原魁[1] 杜清秀[1] 徐春[1] 

机构地区：[1]中国科学院自动化研究所,北京100080

出　　处：《自动化学报》2003年第4期616-621,共6页Acta Automatica Sinica

基　　金：NationalNaturalScienceFoundationofP .R .China (6 0 2 730 2 8);"863"ProjectofP.R.China(2 0 0 1AAN114 2 0 0 )

摘　　要：利用数据手套CAS Glove作为输入设备 ,提出了一种基于模糊神经网络的中国手语单手静态词汇的识别方法 :首先利用经验知识为每个词汇创建模糊规则 ,然后通过学习确定各模糊子集隶属函数中的参数 .对于参数的学习 ,提出了一种适用于分类器的可微经验风险函数 ,该函数能够有效地利用梯度下降法进行最小化 .In this paper, a novel recognition method of single hand static words of Chinese Sign Language based on fuzzy neuro network is introduced. First, the fuzzy reasoning rules and the network structure are established using empirical knowledge. Then the membership function parameters for each fuzzy subset are obtained by learning. For the learning process, a new kind of empirical risk function is proposed which is differentiable and can be minimized by gradient descent strategy. This method is compared with o...

关 键 词：中国手语 模糊神经网络 推理规则 经验风险函数 CAS-Glove 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]