一种应用于超大数据库的人脸识别方法

A Face Recognition Method Applied to Huge Databases

针对数据库中待识别人脸数目巨大,计算困难的问题,提出了一种基于小波变换和主成分分析的人脸识别方法。该方法首先对数据库中的人脸进行小波变换,然后对小波变换后的图像进行主成分分析提取特征,将特征送入BP神经网络进行训练、识别。经过对比实验,结果表明:该方法可以在保证较高识别率的前提下,大大降低数据维数,从而该方法可以应用于超大数据库的人脸识别。

Traditional face recognition methods can't work well,if there are too many faces in the database. In order to solve the problem,this paper presents a face recognition method based on wavelet transform and principal component analysis. Experimental results show that the method can guarantee a high recognition rate,by greatly reducing the data dimension.