An MTCNN-Based Attendance Monitoring System Powered by Inception-ResNet-V1

Face recognition technology is pivotal in modern attendance monitoring systems, leveraging Convolutional Neural Networks (CNN) for high accuracy. Our solution addresses real-world challenges like hardware expenses and appearance variations by processing limited resources through pre-trained models that demand comparatively lower computational power. We employ the Multi-Task Cascaded Convolutional Neural Network (MTCNN) for face detection and the Inception-ResNet-v1 model for face verification and recognition. Our system features a user-friendly web interface for teachers and students, facilitating seamless photo uploads after user authentication and entering respective classroom codes. The models efficiently identify faces and cross-referenced them with pre-existing database embeddings, generating an attendance sheet for the day and a cumulative attendance sheet in.csv format, readily accessible via the website. Our approach that ensures streamlined attendance management has attained a remarkable accuracy rate of 98.13%, emphasizing its practicality and applicability across various environments and devices.