基于小波包和组合分类器的脑电信号分类

发布时间：2018-03-31 02:15

本文选题：脑-机接口　切入点：特征提取　出处：《计算机工程与应用》2016年18期

【摘要】：为了提高脑思维任务分类精度,提出了一种基于小波包分解和多分类器投票组合的运动想象任务分类方法。该方法利用小波包分解对经过预处理的脑电信号进行分解,提取所有频带上的相对小波包能量特征;根据不同脑思维任务下左右半脑各通道间的差异性对C3、C4两通道求取特定频带上的小波包系数的L-2范数作为特征;采用基于投票策略的组合分类器对两种联合特征进行分类,得到了92.85%的识别精度。实验结果表明,联合特征向量较好地反映了左右手运动想象脑电信号的事件相关去同步(ERD)和事件相关同步(ERS)的本质特性;组合分类器识别效果优于单一分类器。
[Abstract]:In order to improve the classification accuracy of brain thinking tasks, a new task classification method based on wavelet packet decomposition and multi-classifier voting is proposed, which decomposes preprocessed EEG signals by wavelet packet decomposition. The relative wavelet packet energy features in all frequency bands are extracted, and the L-2 norm of wavelet packet coefficients in a given frequency band is obtained according to the differences between the left and right hemispheric channels under different brain thinking tasks. The combined classifier based on voting strategy is used to classify the two joint features, and the recognition accuracy is 92.85%. The experimental results show that, The joint feature vector well reflects the essential characteristics of event-related desynchronization (ERD) and event-related synchronization (ERS) of the left and right hand motion imagination EEG signals, and the performance of the combined classifier is better than that of the single classifier.
【作者单位】：中国地质大学(武汉)机械与电子信息学院;
【基金】：湖北省自然科学基金项目(No.2012055077) 中央高校基金项目(No.2012079108)
【分类号】：R741.044;TN911.7

【相似文献】