基于MobileFaceNet改进的人脸识别课堂签到系统

Face Recognition Classroom Sign-in System Based on an Improved MobileFaceNet

随着科学技术的不断发展,各类智能设备在人们的生活中随处可见。建立了一类人脸识别课堂签到系统,该设备基于深度学习算法进行人脸识别,通过提出的阶梯融合模块对基础人脸识别网络MobileFaceNet进行改进,使其拥有更高的识别准确率,该网络在LFW、AgeDB-30和CFP-FP数据集上均取得了较好的结果。接着,建立了签到数据库用于记录受试者的基本信息和签到信息,使得课堂签到过程的数据更具实时性,同时更便于教师管理。该系统的建立将原先繁琐的人工签到任务智能化,减少人力资源,使得课堂签到任务更加高效和准确。

This paper establishes a class of face recognition classroom check-in system,which is based on deep learning algorithm for face recognition,and improves the basic face recognition network MobileFaceNet by the proposed step fu-sion module to make it have higher recognition accuracy,and the network achieves better results on LFW,Age DB-30 and CFP-FP datasets.Then,a check-in database is established to record the basic information of the subjects and the checkin information,which makes the data of the classroom check-in process more real-time and easier for teachers to manage.The system is designed to reduce human resources and make classroom check-in tasks more efficient and accurate by in-telligentizing the previously tedious manual check-in tasks.