Research On Face Recognition Related Issues In Access Control Systems

With the progress of human society and the development of computer technology,in recent years,especially the research of entrance guard system based on biological characteristics is pursued by many researchers.However,within lots of biological characteristics,the entrance guard system which is characterized by face has become a hot topic studied by researcher and research institutions at home and abroad.Compared with other human biological characteristics,face can be easily accepted by users for its convenience.The face recognition technology has gained a great achievement after many years’ joint efforts,but it still has difficult problems which remain to be solved.In a complex application scenario,the interference factors of the background(such as: the face image on the wall,the red curtains,doors and Windows and so on)will lead to the undetected error rate’s increase of the face;The uncertainty of the facial features as well as a variety of expressions and poses will directly affect the recognition effect.In order to overcome the influence of background interference factors in the complicated application scenarios,we first detect moving objects detection of video sequences,then detect YCbCr color model based on moving target detection,finally we pinpoint human face through morphological processing,regional markers,and horizontal projection,thus it can effectively eliminate the interference of background factors.In the face recognition phase,the PCA can extract face features and also reduce the dimension of feature extraction so that it reduces the computational complexity.Therefore,it can be used for face recognition when combining PCA with SVM.The experimental results show that the method can recognize faces in the right time.