多角度人脸检测与识别方法研究

摘要

In this paper,a novel multi-view face detection (MVFD) algorithm based on improved Haarlike features and cascade AdaBoost classifier and face recognition algorithm based on OpenCV is presented.Through the analysis of the related theory of PCA and 2DLDA,combined with the advantages design and implement a face recognition algorithm for MVFD.At the same time its performance weakly dependent of angle information of face recognition.The experiments showed that an improved method in face recognition is achieved by using the additional multi-view information.%本文研究一种多角度人脸检测与识别方法,采用一种改进的Haar-like特征,调用OpenCV计算机视觉库实现人脸的多角度检测和人脸图像的存储.通过分析PCA和2DLDA的相关理论,结合两者优势设计并实现一种多角度人脸识别方法,提出了一种解决人脸角度因素对人脸识别系统影响的方法.