基于混合SVM与AdaBoost分类的面部表情识别的人机交互

基于二维主元分析（2DPCA）方法提出一种混合支持向量机（SVM）与AdaBoost算法的面部表情分类方法。首先，该方法对灰度图像进行人脸检测，通过小波变换和二维主元分析得到特征数据，有效地减少了计算量；然后，采用SVM方法对特征数据进行分类学习，得到初始分类器；最后，通过AdaBoost算法对SVM分类结果进行进一步加强，形成强分类器，提升了分类能力，确保了表情识别工作，并实现基于面部表情识别的智能轮椅的人机交互的鲁棒性。实验结果表明：该方法不仅有效地提高了样本的分类能力，而且降低了计算的复杂度，在智能轮椅人机交互实验中的平均识别率达到92.5%。

Human computer interaction technology of expression recognition based on SVM and AdaBoost classifier

This paper is based on the 2DPCA method and introduces a facial expression classification on method which is based on a SVM and AdaBoost algorithm. Firstly, using the method for the face detection of the gray image and obtaining the characteristics data through the wavelet transform and 2DPCA, We reduced the amount of computation effectively. Secondly, we obtain the original classifier by the SVM method to classify learning characteristics data, then through the AdaBoost algorithm to further strengthen the SVM classification results, forming the strong classifier and improving the classification ability. This ensure finishing the work expression recognition and realizing the robustness of the man machine interaction based on the facial expression recognition of the intelligent wheelchair. The experimental results show that this method not only effectively improves the classification ability of the sample, but also reduces the computational complexity, with an average recognition rate of 92.5% in the intelligent wheelchair human computer interaction experiment.