基于EOG的眼动信息提取与分类研究

Research on information extraction and recognition of the eye movements based on EOG

基于眼电(Electro-oculogram,EOG)的人机交互系统(HCI)是生物电信号处理领域的研究热点之一。在研究眼动信息的基础上,提出了一种EOG扫视信号特征提取与分类算法,该算法提取扫视信号的线性预测(Linear Predictive Coding,LPC)系数,对其作差分运算获取一阶差分线性预测系数,与归一化极值作为组合特征参数,通过神经网络对样本信号分类。实验室环境下,采用所提该法对来自6名眼部功能均正常的受试者扫视样本分类,平均分类正确率超过92%。实验表明,该法能准确地描述EOG扫视信号,具有较高实用价值。

The research on human-computer interface （HCI） based on EOG is a hotspot in the field of bio-signal processing.Research on the basis of eye movements information,this paper presents a feature extraction and classification algorithm based on EOG saccadic signals,which extracts the saccadic signals＇ linear predictive coding （LPC） coefficients and then extracts first-order differential linear predictive coding coefficients and normalized extremum as the characteristic parameters.The sample signals are classified by an artificial neural network （ANN）.Under the laboratory environment,the samples,got from six different subjects with normal eye function,are classified using the proposed algorithm,the average classification accuracy rate reaches more than 92%.The results show that the proposed algorithm can depict the EOG saccadic signals＇ features accurately and have practical use.