Research And Implementation Of Class Attendance System For Mask Face Recognition

Class attendance has always been completed by the teacher by name.This method is inefficient and low in information.It is impossible to avoid the phenomenon of students being absent for no reason,answering on behalf of others,etc.,so that the teacher cannot accurately grasp the student’s class attendance.Face attendance technology,as one of the research hotspots in recent years,has gradually been promoted and popularized in campuses and enterprises,and has received unanimous responses from campus teachers,students,and corporate personnel.We know that face recognition technology under constrained conditions has been widely used in social life.However,many face collections in natural environments are in unconstrained environments,such as large-area occlusion,poor lighting conditions,or people.If the angle of the face pose is too large,it will destroy the extraction of facial features,thereby reducing the accuracy of face recognition.In the context of epidemic prevention and control,research on the application of this technology in campuses has become particularly urgent and important.This paper compares the face recognition strategies of masks in the classroom attendance system.After understanding the research trends of related technologies at home and abroad,finally adopts the face detection and face recognition algorithms based on Convolutional Neural Network(CNN).The main work of this paper is as follows:(1)The application status of classroom attendance system and face recognition technology is studied,the traditional methods and deep learning methods used in face detection and recognition are compared,and the model based on convolutional neural network is introduced for the problems they face.Structure to carry out the superiority of face detection and recognition.(2)Aiming at the characteristics of simultaneous detection of multiple faces required for classroom attendance,this paper uses a face detection model composed of three cascaded convolutional neural networks to detect and align multiple faces in real time.In view of the low-quality image problem in the unconstrained environment,the face image quality evaluation function is used to score the input face image,so that the high-scoring and highquality image directly enters the next face recognition process.The low system is directly screened out and re-detected,which improves the performance of system identification.(3)At the same time,this paper uses the Face Net model to learn the features of the face image.Face Net maps the face image to the Euclidean space,measures the similarity of the face through the spatial distance,and uses a triple loss function to make the model learning more capable Efficient.(4)Based on the basic process of the classroom attendance system,according to the basic architecture design and function design of the system and database design,etc.,using Python language and Django framework,combined with My SQL database,a mask face recognition classroom attendance was finally designed and implemented.system.Experimental results show that this system has good multi-person real-time detection and face recognition performance,and also has good robustness and accuracy for unconstrained environments.