基于信号投影能量特征的脑电意识动态分类  

Dynamic Motor Imagery Classification with Signal Power Projection based Feature

作　　者：王永轩[1] 邱天爽[1] 刘蓉[1] 李春月[1] 马征[1] 

机构地区：[1]大连理工大学电子信息与电气工程学部,大连116024

出　　处：《信号处理》2012年第8期1059-1062,共4页Journal of Signal Processing

基　　金：国家自然科学基金资助项目(60940023;61172108;61139001;61005088);国家科技支撑计划项目(2012BAJ18B06);机器人技术与系统国家重点实验室开放研究项目(SKLRS-2010-ZD-07)

摘　　要：针对脑电意识任务动态分类问题,本文提出了一种基于投影能量的特征提取方法来提取反映不同思维状态的脑电特征,并结合信息累积后验贝叶斯方法进行分类以提高脑-机接口系统的分类正确率。该方法通过使两类信号在投影基上的平均投影能量比达到极值,从而达到提高脑电信号分类准确度的作用。实验结果表明两个运动想象数据集上的最大正确率都达到90%左右,最大分类准确率、kappa系数和最大互信息等评价指标的比较也表明该方法能够有效提高BCI系统的性能,具有较好的实用性。The brain-computer Interface （BCI） gives interactive communications between people and the machine, and has fascinated the researchers over the last couple of years. However, the BCI system suffers from a low information trans- mission rate, low accuracy and poor interactive performance, which is the bottleneck for the promotion of BCI-actuated sys- tem. Therefore, to classify different motor commands fast with minimal error is an important problem in the BCI system. For the dynamic classification of motor imagery mind states in the brain-computer interface （ BCI）, we proposed a power projec- tion based feature extraction method to classify the EEGs by combining information accumulative posterior Bayesian ap- proach. This method improves the classification accuracy by maximizing the average projection energy difference of the two types of signals. The experimental results on two motor imagery datasets show that the maximum classification accuracy is a- bout 90%. With three indexes, i.e. maximum classification accuracy, kappa coefficient and mutual information, the effec- tiveness of this method is demonstrated.

关 键 词：脑-机接口 运动想象 投影能量 贝叶斯分类 

分 类 号：TN911.7[电子电信—通信与信息系统]