支持向量机方法在运动意识识别中的应用

Application of support vector machines in the recognition of motor intention

为有效识别与运动想像相关的脑电模式,提出基于支持向量机(SVM)的运动意识分类新算法,利用sym2小波基函数对脑电(EEG)信号进行6尺度分解后,从每级分解中提取绝对值最大的小波系数作为信号特征,构成有效特征向量输入SVM分类器,实现基于EEG的运动想像模式识别.实验数据采用脑机接口竞赛(2003)的脑电数据,实验结果表明采用径向基核函数的SVM分类器可有效地对EEG进行运动想像分类,具有良好的泛化推广能力,为脑机接口的运动意识分类提供了新思路.

The extraction and recognition of electroencephalogram （EEG） features encoding motor imaginary can help the disabled user due to severe paralysis to control the cursor movement or other actuators. A novel algorithm based on support vector machine （SVM） to classify different EEG patterns of motor imagery was proposed, in which the EEG signal features was extracted using wavelet transform （WT） and the number of decomposition levels was chosen to be 6. The maximum of the wavelet coefficients in each sub-band were used to serve as the inputs of SVM. Experimental results show that SVM can classify different EEG patterns of motor imagery effectively.