多角度不同表情下的人脸识别

Face Recognition under Different Expressions and Multi-views

多角度及不同表情下的人脸识别是人脸识别领域的一个难题。本文将二维主元素分析法与贝叶斯判据相结合设计了多角度不同表情下的人脸识别算法。首先,利用二维主元素分析法计算人脸的特征矢量空间,并将训练集和测试集中的数据向该特征矢量空间进行投影,然后使用贝叶斯判据进行识别。该方法集中了二维主元素分析法计算简单、速度快及统计分类器识别率高的优点。实验结果显示,该方法计算简单,对具有表情变化及不同角度的人脸的识别率高。

Face recognition under different expressions and multi-views is a difficult problem. A new algorithm for face recognition under different expressions and multi-views is presented in this paper. Firstly, the 2D PCA algorithm is used to compute the eigenvector space of the fact And then, the faces in the training set and testing set are projected to this face spac. Secondly, Bayes rule is used to design the classification designer. The advantages of simple computation and quick speed of two-dimensional PCA and the high recognition rate of statistical classification are combined into the new method. The experimental result shows that the method introduced in this paper has the advantages of simple computation and high recognition rate under different expressions and multi-views.