Development of a Face Recognition-Based Attendance System
Chandu Vaidya, Sonam Chopade, Prashant Khobragade, Kalpana Bhure, Aditya Parate · 2025
Facial recognition-based attendance systems are all set to revolutionize the old ways of taking attendance, providing accurate nature and ease of in handling. The purpose of work is to present an advanced system that works with the CNN capabilities of the dlib library to ensure robust, real time face recognition with an accuracy rate above 96%. The system is designed to be effective with a standard camera configuration and does not require specialized hardware. The methodology is to capture the facial image of people entering a defined region and then filtering them using deep learning-based CNN models to locate some unique facial features. During the first time enrolment we securely encode and store these features. During operation, the system compares live facial input to pre-registered templates, marking attendance along with timestamps in real time. The system is shown to be efficient and adaptable across multiple scenarios and high precision. In conclusion, the results emphasize that complex machine learning incarnations, particularly those using dlib CNN convolutional models, are effective for boosting the effectiveness, scalability, and overall practicality for deploying facial recognition-based attendance solutions in wide scale.