融合原始样本和虚拟样本的人脸识别算法

由于有限的存储容量和捕获图片的时间,实际的人脸识别系统往往只能获得少量的训练样本,但是,在小训练样本情况下大多数人脸识别算法都会遇到困难.因此,为了提高人脸识别的分类正确率,提出了一种融合原始样本和虚拟样本的人脸识别方法.该方法先利用人脸的对称性来构造虚拟训练样本;然后,利用协同表示方法分别对原始训练样本和虚拟训练样本进行分析,并且分别得到每一类训练样本的重建误差;最后,将原始训练样本和虚拟训练样本的同一类重建误差进行加权融合并得到最终的分类结果.大量的实验结果比较分析表明,该方法可以获得更好的识别效果.

Face recognition based on the fusion of original sample and virtual sample

A real face recognition system often can obtain only a small number of training samples because of the limited storage space and the limited time of capturing images. However, most face recognition algorithms will encounter difficulties in the case of small training samples. Therefore,we propose a new face recognition method based on the fusion of original sample and virtual sample to improve the classification accuracy of face recognition. First, the proposed method exploit the symmetry of the face to generate virtual samples. Then, it use the collaborative representation method to perform the original and virtual training samples classification, respectively and it will obtain reconstruction error of every class. Finally, it combines the reconstruction error of the same class obtained by the original and virtual training samples to conduct weighted fusion and gets the ultimate classification result. A large number of experimental results show that the proposed method can obtain better recognition effect.