The Design And Implementation Of Classroom Attendance System Based On Video Image

With the evolution of society,attendance plays a more and more important role in the management system,especially classroom attendance.The traditional way of attendance is calling names manually,and record the attendance situation on the paper.With the rapid development of the computer industry,there have been a variety of attendance software,attendance system,but the manual form is still the main way of student attendance,due to the cost of deployment problems.Teachers need to take the class time to check the attendance of students,which is a waste of time.So,the accuracy and efficiency of the attendance is to be improved by this way.An attendance system in classroom time based on video images is designed and implemented in this paper.Firstly,the research status of face recognition at home and abroad is analyzed,and the development process of the attendance system is expounded.Then the classic algorithms related to face detection and face recognition are introduced,and the advantages and disadvantages of these methods are analyzed.Then,the key technology SSD [1] of this paper is introduced in detail,and face detection is realized by this method,and VGGFace [2] and LMNN [3] are used to realize face recognition.The main work of this paper is as follows: Firstly,a face detection model that suitable for classroom attendance scene is trained by SSD in the existing face dataset CelebA [4].And then the face feature is extracted with VGGFace depth network.In this paper,a tagged face image dataset is created manually,and a face classifier is trained with LMNN algorithm on the dataset,the face detection model and face classifier are applied to the classroom attendance system finally.The face detection and face recognition process are realized by using the depth neural network,which improves the precision of detection and recognition,and adopts the method of cascade network to improve the system performance.The classroom attendance system designed in this paper is based on the video images,through the continuous video frame for face detection and recognition to achieve the purpose of attendance.This fully automated design makes classroom more efficient.