SVM和小波包变换在动作模式识别中的应用

支持向量机（SVM）是一种线性机器，广泛用于模式分类和非线性回归。对于很多低维非线性可分的模式，如果我们能够提取合适的高维特征向量，则模式往往在高维特征空间是线性可分的。本文利用小波包变换提取动作的特征向量，将各种动作信号映射到特征空间形成一定维数的特征向量，然后采用SVM进行动作识别。试验证明。当特征空间维数合适时，利用SVM进行动作识别效果良好。

The Application of the SVM and Wavelet Packet Transformation in Motion Recognition

The support vector machine(SVM) is a linear classification machine,it is used commonly in the pattern recognition and nonlinear regression.Many non-linear classification problems,where we extract some advisable dimension vectors to form a characterization space,can be solved as linear classification problems.Using the wavelet package to extract advisable characterization vectors from EMG signals as SVM input vectors,this paper works on solving the motion pattern classification.The experiments proved this method has good performance.