An improved non-perceptual VIPLFaceNet face recognition algorithm for classroom attendance system

Aiming at the low detection rate of the existing classroom attendance system and the inconvenience of data query, a non-perceptual classroom attendance system based on face recognition is proposed and designed. Using the Android development platform, the image collected by the front-end was first transferred to the server through OkHttp3 technology. Then the information of the class in the database MySQL was retrieved. Then the face image of each student was filtered through the improved Fust face detection algorithm, and the similarity value within the class and similarity value between the class generated VIPLFaceNet face recognition threshold, which recognized the screened face images and obtained the attendance result. Finally, the attendance result was sent to the front end, the administrator could access the server to query attendance data. The experimental results showed that the recall rate of the improved Fust face detection algorithm and the recognition rate of the VIPLFaceNet face recognition algorithm could reach 90.18% and 98.79% respectively.