文中利用主成分分析提取图像的特征信息,然后将此特征数据作为分类器的输人数据。文中采用的分类器——支持向量机是一种能在训练样本数很少的情况下达到很好分类推广能力的学习算法。再利用二叉树判别策略来对多类人脸图像进行了识别,实验取得了较好的效果。
At first,makes use of PCA to extract the feature information of images,then makes the feature data to be the input information of classifier. Adopts a classifier support vector machine which is a very good training algorithm,which can acquire very good generalization when the training datum are very few. At last, recognize the multi - images with binary tree and acquire very good results.