Font Size: a A A

Technology And Research Of Real-time Face Recognition System Based On PCA

Posted on:2022-05-27Degree:MasterType:Thesis
Country:ChinaCandidate:R J PengFull Text:PDF
GTID:2518306530980209Subject:Electronics and Communications Engineering
Abstract/Summary:PDF Full Text Request
With the development of the society,the technology of automatic identification to protect personal safe and information safe is required by the trend of the times.Compared with fingerprint recognition,face recognition as an authentication technology can be performed without touching.The accuracy,convenient,safety and other characteristics of the Face recognition have attracted internal and external scholars.Face recognition has been applied in home security access control,video surveillance,intelligent transportation,and medical care.The PCA is a superior algorithm in face recognition that has achieved excellent results in data dimensionality reduction and face feature extraction.this article takes the laboratory access control system as the research background,designs and implements a face recognition system which based on KPCA+LDA+SVM algorithm.The main research contents of this paper are as follows:(1)Aiming at the problem that the PCA algorithm has a low recognition rate of face images,cannot handle the non-linear features of the face and the unused category information,this paper proposes the KPCA+LDA+SVM face recognition algorithm.Based on the original PCA algorithm,this algorithm proposes a fusion of kernel principal component analysis(KPCA)and linear discriminant method(LDA)to extract features of face images,and then combines support vector machines(SVM)to classify and recognize face images.The analysis results show that the effect of the algorithm proposed in this paper has been improved.(2)Establish a laboratory staff registration system interface to complete the registration of laboratory staff.On this basis,a personal face database(self-built face database)was established.The database had a capacity of 400 faces which is composed of 10 facial angle of 40 people.The effectiveness of the KPCA+LDA+SVM algorithm proposed in this article has been experimentally verified in the ORL face database,Yale face database,and self-built face database.The obtained face recognition rates are 98.13%,96.00%,98.12%.(3)Build a laboratory face recognition system based on the improved KPCA+LDA+SVM algorithm,and simulate the system in MATLAB,increase its visualization function,and present it with a GUI human-computer interaction interface.Finally,the functions of real-time face recognition,local video recognition and local photo recognition of laboratory personnel are realized.The system can correctly recognize the human face,which shows that the research of the system is feasible and has certain reference value for practical research.
Keywords/Search Tags:Face recognition, Principal component analysis, Feature extraction, Support vector machine, Laboratory access control system
PDF Full Text Request
Related items