Deep Learning-Based Automated Attendance for Blended Environments with Real-Time Synchronization

In the post-pandemic era, the use and demand for blended learning platforms have grown significantly, leading to a diverse mix of online and offline classes. As a result, teachers face challenges in managing attendance records for both types of classes, which hinders the efficient and seamless integration of learning experiences. To address this issue, we have developed AutoAttendance, a comprehensive attendance management system that automates the process of capturing and merging attendance data from multiple online class platforms and offline face-to-face classes. The system integrates five popular online class platforms i.e zoom, google meet, etc., and employs a FaceNet deep learning model to accurately detect and recognize students in still images or live video streams for offline classes. A unique methodology, utilizing default_id and platform_id, is implemented to merge attendance records from various sources seamlessly. This demonstrates an impressive accuracy of 99.63% for offline attendance tracking using the FaceNet model and 100% accuracy for online classes, offering a reliable and efficient solution for automating attendance management with real-time synchronization in blended learning environments.