眼电信号是人机交互系统中的一种重要的信息源,该文提出了一种眼电信号特征提取与分类算法。首先研究小波包变换,提出以小波包分解系数作为眼电信号特征,通过支持向量机进行分类识别。在实验室环境下,采用该方法对6名眼部功能正常测试者的样本数据进行分类,平均识别率达到96.83%,具有很高的实用价值。
EOG signal is one of the important information sources in HCI system, the feature extraction and classification algorithm of EOG signals is put forward. The wavelet packet transform is studied firstly, and then proposing the idea of taking the wavelet packet coefficients as the feature of EOG signals, and the sample signals are classified by support vector machine (SVM). Under the laboratory environment, the sample data collected from six students with normal eye function are classified. The average rate of identification reaches 96.83 %, which proves it has high practical value.