RESEARCH ON ATTENDANCE SYSTEM OF FACE RECOGNITION

Abstract

With the continuous development of computer vision technology, face recognition technology is widely used in identity authentication system. The main research content of this paper is the application of face recognition technology in the swipe card attendance system. We all know that most of the traditional swipe card attendance system is to swipe the electromagnetic card to sign in, which inevitably leads to the phenomenon of signing on behalf of others and forgetting to bring the magnetic card. In order to solve the inconvenience of traditional attendance system, in this paper, we focus on the research of a face recognition system of swiping card attendance. We only need to brush the face to punch in. Considering the disadvantages and low accuracy of traditional pattern recognition methods, this paper adopts deep learning face recognition technology. Firstly, through tensorflow deep learning framework, a separable convolutional neural network is built; secondly, face data sets are trained and classified, and relevant parameters are set to train a more suitable face model; finally, MFC visual programming integration process is used in order to realize the computer-side face recognition attendance system. The accuracy of face recognition technology in this paper can reach 98% in 1-to-1 recognition, which can basically meet the needs of attendance system.