Multi Face Recognition based Attendance Marking System using MTCNN with FaceNet

The document describes a face recognition system for Attendance Management which enhances security alongside operational efficiency in guard log records. The attendance tracking methods based on manual procedures and RFID cards facilitate proxy incidents and generate operational delays. A standard camera through the combination of MultiTask Cascaded Convolutional Neural Network (MTCNN) with FaceNet enables the proposed system to automate attendance recording without needing extra equipment. Three major functions form the basis of this system: training sets up the parameters and both attendance tracking and student identification access. Training the system required 908 images from 21 students and produced 100% face detection accuracy and 86% recognition accuracy during testing of 108 images. The system operates at low cost with touchless operation while defending against fraudulent attendance practices. Recognized improvements in development focus on precision enhancements when dealing with images that are not frontal while developing web and mobile platform interconnectivity.