论文部分内容阅读
为了提高基于眼电图(EOG)的扫视信号识别正确率,提出了一种基于共同空间模式(CSP)的扫视信号特征提取算法.该算法首先利用事先标注好的标签数据进行CSP空域滤波器设计,并采用联合近似对角化的方法解决多分类问题;在此基础上,使用该滤波器对原始多导联眼动信号进行空域滤波,滤波输出即为扫视信号的特征参数.在实验室环境中使用支持向量机对上、下、左、右四类扫视信号进行识别,所提算法的平均正确率达到了97.7%.实验结果表明基于CSP的扫视信号特征提取算法在眼动信号分析中呈现出良好的分类性能.
In order to improve the accuracy of EOG-based glance signal recognition, a feature extraction algorithm based on the common spatial mode (CSP) is proposed, which first uses the pre-labeled tag data for CSP spatial filter design , And the method of joint approximation diagonalization is used to solve the multi-classification problem. On this basis, the filter is used to filter the original multi-lead eye-moving signal in spatial domain, and the output of the filter is the characteristic parameter of the saccade signal.In the laboratory environment , The left and right scanning signals are identified by using SVM, the average accuracy of the proposed algorithm is 97.7% .Experimental results show that the feature extraction algorithm based on CSP is presented in the eye movement signal analysis A good classification performance.