基于小波和神经网络的视觉诱发电位识别方法  

Recognition method of visual evoked potential based on wavelet & neural network

作　　者：何庆华[1] 吴宝明[1] 彭承琳[2] 王禾[1] 钟渝[1] 

机构地区：[1]第三军医大学大坪医院野战外科研究所创伤,烧伤与复合伤国家重点实验室,重庆400042 [2]重庆大学生物工程学院,重庆400044

出　　处：《仪器仪表学报》2007年第6期1003-1006,共4页Chinese Journal of Scientific Instrument

基　　金：国家自然科学基金(30300418);重庆市自然科学基金(CSTC;2005BB2187)资助项目

摘　　要：在基于瞬态视觉诱发电位的脑机接口研究中,通过视觉诱发电位信号的提取与识别产生脑机接口控制信号。采用累加平均和小波滤波提取强噪声背景下微弱的视觉诱发电位。在小波变换域求取特征向量,将特征向量输入感知器进行视觉诱发电位模糊识别,产生脑机接口控制信号。实验表明,小波变换域特征向量提取方法能有效地实现信号的去噪、降维和特征提取,基于神经网络的模糊识别算法能比较准确地识别视觉诱发电位,有利于提高脑机接口的通讯率。In the research of the transient visual evoked potential （VEP） based brain-computer interface （BCI）, VEP signal extraction and recognition are essential for obtaining BCI control signal. Wavelet filtering and averaging method are used to extract the weak VEP signal from strong background noise. The feature vectors obtained in wavelet transform domain are input to the perceptron to achieve the fuzzy recognition of VEP and produce brain-computer interface control signal. Experiments show that the feature extraction in wavelet transformation domain can effectively extract VEP features, reduce noise and decrease the dimensionality at the same time, and the neural network based VEP fuzzy recognition algorithm can recognize VEP signal correctly, which is useful to increase the information transfer rate of brain-computer interface.

关 键 词：视觉诱发电位 脑机接口 小波变换 神经网络 

分 类 号：R318[医药卫生—生物医学工程]