AI-Based Face Recognition System Using Dlib for Real-Time Attendance

This report describes an artificial intelligence based real time facial recognition system, implemented using the open source Dlib library, to assist in attendance tracking. Traditional systems for managing attendance (manual register and biometric) are labor intensive and prone to errors. One common error is proxy attendance. Computer vision techniques can be utilized to automatically identify a person in real time using facial detection and recognition. Facial landmarks and facial encoding within the Dlib library will be used by the system to facilitate facial recognition. Image processing will be performed using OpenCV. The system will create a 128-dimensional vector for each facial feature extracted from images. Each vector will be matched with an existing encoding to identify a person. When a person is successfully identified by the system, the time and date will be recorded and attendance will be marked in real-time. In an experimental environment, the system successfully identified individuals with high accuracy. Manual effort was also significantly reduced. This technology will provide educational institutions and organizations with a contactless, reliable, efficient and scalable attendance management system.