带有年龄信息的PCA人脸识别技术  

PCA Face Recognition Technology with Age Information

作　　者：陈文兴 岳靖 付浩 CHEN Wen-xing, YUE Jing, FU Hao(School of Mathematical Statistics, Ningxia University, Yinchuan 750021, Chin)

机构地区：[1]宁夏大学数学统计学院

出　　处：《数学的实践与认识》2018年第6期118-130,共13页Mathematics in Practice and Theory

摘　　要：基于主成分分析（PCA）的人脸识别算法．其中包括K—L变换、SVD分解、特征脸的构建、人脸年龄估计和人脸识别等过程．使用了Manhattan、Euclidean、Cosine这3种距离进行判别，实验得出Euclidean最优，最后，用MATLAB编程计算出人脸识别的准确率．结合柔性模型（AAM）和分层训练样本实现带有年龄信息的人脸识别技术；同一年龄段内训练样本数越多，识别率越高．This paper introduced an algorithm of face recognition which is based on principle component analysis（PCA）. It Include the K-L transformation, （SVD）decomposition. What＇more, it also have the biulding of feature face, estimation of age and the processing of the face recognition. To compare and judge the result, we use the minimum distance method. Such as Manhantan, Euclidean and Cosine distance to calculate and simulate it by using MATLAB to get the accuracy of PCA face recognition. Finally, we combined the flexible model with stratifying the samples to train to get the technology of face recognition with age information. The more the number of training samples in the same age group which has a higher the Haman rate.

关 键 词：人脸预处理 PCA人脸识别 柔性模型 特征脸 K—L变换 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]