摘要
针对实时人脸识别易受光照变化影响的问题,提出了一种基于Adaboost人脸检测、局部二值模式(Local binary patterns,LBP)特征提取、PCA(主元分析Principal component analysis)特征降维和SVM(支持向量机Support Vector Machine)特征分类等人脸识别方案,经过长期的对人脸识别技术的研究在多个人脸数据库上进行了实验,结果表明该文算法对光照具有较好的鲁棒性,提高了执行速度和识别率。
In this paper,a new face recognition method combining local two-valued model with embedded hidden markov model is proposed to solve the problem of real-time face recognition. This method firstly pretreated the input face image and then extracted the feature vector,then the extracted feature observation vector was carried into the EHMM for training or recognition. The experiment results show that the algorithm has better robustness and better recognition rate.
作者
顾玮
GU Wei(Xuzhou Higher Normal School Xuzhou 221116)
出处
《办公自动化》
2018年第2期41-43,共3页
Office Informatization
关键词
人脸识别
LBP
PCA特征降维
SVM分类识别
Face recognition LBP Characteristics of PCA SVM classification identification