Research And Implementation Of Face Recognition Based On PCA And SVM

Face recognition is one of the most popular topics in the field of scientific research.Its application has penetrated into various fields such as financial system,information security,public safety and so on.In the horizontal comparison,face recognition has obvious advantages in many pattern recognition methods,and has natural advantages,such as naturalness,non-intrusive,low acquisition cost and strong human-computer interaction,and the application scene is more extensive.Vertical analysis,the major listed companies have introduced a large number of commercial face recognition products,face recognition will show explosive growth in the future.Therefore,it is of great practical significance to study face recognition.This thesis focuses on the three aspects of face preprocessing,face feature extraction and face recognition.In particular,the principal component analysis(PCA)and the support vector machine(SVM)algorithm are studied deeply,and the linear discriminant analysis(SVM)is not introduced for the lack of classification information.(LDA),an improved framework for face recognition based on PCA and SVM is proposed,and a real-time face recognition system is implemented based on the improved framework.The main work of this thesis has the following five parts.(1)Researched a variety of commonly used face image pre-processing techniques,including gray-scale color image,gray level transformation,histogram equalization and image geometric normalization.The face image pre-processing technology reduces the influence of external factors such as illumination,attitude and shooting angle on the image,and realizes the standardization of the image,which lays a good foundation for the subsequent face recognition work.Finally,used Matlab platform to do the relevant algorithm test work.(2)Researched the theoretical basis and specific implementation process of face recognition based on PCA in detail.Since the calculation of eigenvalues and eigenvectors of the covariance matrix using PCA algorithm directly are too much,the SVD theorem is introduced to realize the indirect operation of the principal subspace.And a lot of experiments are carried out by using Matlab.The advantages and disadvantages of PCA algorithm are analyzed.It is concluded that PCA algorithm has good reduction effect,but there is no classification effect.(3)Researched the theoretical basis of linear support vector machine in detail,and the advantages of SVM classifier in small sample and learning are discussed.Then the kernel function is used to extend the linear support vector machine,so that the support vector machine can meet the classification requirements of face recognition.Discussed the implementation of multi-classification SVM method.(4)Aiming at the problem that the classification information is not used in the PCA algorithm,the linear discriminant analysis is introduced and the improved face recognition framework based on PCA and SVM is proposed.Finally,the face recognition framework of PCA + LDA + SVM is formed,On the Matlab platform,a lot of experiments have been carried out to discuss the influence of the number of dimensions and the number of training samples on the algorithm,and verify its good classification effect.(5)Based on the improved face recognition framework,a real-time face recognition system is realized with Visual Studio 2010,Open CV library and Qt framework.Through the dynamic face recognition test,the recognition rate is 97.3%,the recognition effect is excellent,the recognition time is 356 milliseconds each time,meet the real-time requirement.