Analysis Of Face Recognition Technology And Its Application In Attendance System

With the progress of science and technology,the use of human face,iris,fingerprint,finger vein and other biological characteristics of the technology gradually replaced the identity card,password and other traditional ways to identify.With the help of biometric technology can effectively improve the user information security and authenticity.In these technologies,face recognition is the most rapid development,at the same time,it also has many advantages,such as collecting face images convenient and quick,do not directly contact the machine,precision identification and so on.Therefore,face recognition in the verification of identity,protection of user information and so have a very broad application space.The core technology of face recognition is able to extract the key features of human face different from others,and these features are the prerequisite and key to the successful realization of efficient face recognition algorithm.In this paper,the principle and performance of the local binary model algorithm and the convolution neural network algorithm are described and analyzed in detail.Then,a mobile attendance system is designed and implemented by combining the two algorithms with the mobile phone positioning technology respectively.The mobile attendance system ensures the validity of the check-in through the dual technology of face authentication and mobile phone positioning.Mobile attendance system consists of mobile client and PC server,client and background server implementation using java language in Eclipse and My Eclipse platform development completed.The function modules of the time and attendance system include face image entry,latitude and longitude sign and check out(using face recognition and positioning technology combination),the user has something leave,IP precision punch and check the month sign time.Attendance module design is reasonable,attendance rules are simple,user operability is strong.Through a large number of experimental analysis,it is shown that the LBP algorithm has the characteristics of high precision,short response time and strong real-time performance.Although CNN algorithm is very high accuracy,but because of the hardware configuration causes CNN reaction speed is relatively slow.