摘 要：人脸识别技术作为一种生物识别技术得到了广泛的应用，实现人脸识别的方法多种多样，其中特征提取是人脸识别最重要的步骤之一。本文重点介绍了两种主流的人脸识别技术，即主成分分析和线性判别分析，具体介绍了这两种方法的工作原理和实现步骤，并通过分析其工作原理说明了它们的优点和缺点。针对这两种方法从识别速率、识别准确率和对各种噪声的鲁棒性等方面进行比较，说明两种方法的最佳使用条件。最后提出当前人脸识别面临的巨大挑战和未来前进方向。
关键词：人脸识别；特征提取；主成分分析；线性判别分析
中图分类号：TP391.41 文献标识码：A 文章编号：2096-4706（2019）24-0100-02
Comparative Study of Two Face Recognition Technologies
QIU Huamin
（Quanzhou Electric Power Skills Research Institute，State Grid Fujian Electric Power Co.，Ltd.，Quanzhou 362000，China）
Abstract：Face recognition technology is widely used as a biometric technology. There are many ways to achieve face recognition.Among them，feature extraction is one of the most important steps in face recognition. This article focuses on two mainstream face recognition technologies—principal component analysis and linear discriminant analysis. The working principles and implementation steps of the two methods are introduced in detail，and their advantages and disadvantages are explained by analyzing the working principles of the two methods. The two methods are compared from the recognition rate，recognition accuracy and robustness to various noises，and the best conditions for using the two methods are explained. Finally，the huge challenges faced by current face recognition and the way forward are proposed.
Keywords：face recognition；feature extraction；principal component analysis；linear discriminant analysis
参考文献：
[1] 蒋小军. 人脸识别技术应用探究 [J]. 现代信息科技，2019，3（14）：91-92.
[2] 孟德顺.PCA 应用中的几个问题 [J]. 西北林学院学报，1995（1）：89-94+42.
[3] 崔自峰，吉小华. 基于线性判别分析的特征选择 [J]. 计算机应用，2009，29（10）：2781-2785.
作者简介：丘华敏（1987.05-），男，汉族，福建上杭人，本科，工程师，研究方向：信息安全、计算机系统应用。