Automated attendance management system using Face Recognition (CNN)

Abstract

This thesis explores the development and evaluation of an automated attendance system using Face Recognition powered by Convolutional Neural Networks (CNN). The system aims to enhance accuracy and efficiency in attendance marking by recognizing and verifying individuals in real time. Trained on a diverse dataset, the CNN model demonstrated high accuracy, precision, and recall in testing, showing robustness in various lighting conditions and angles.Real-world testing confirmed the system’s efficiency with rapid recognition times and low error rates. Compared to traditional methods, the system significantly improved accuracy and reduced manual errors. The findings suggest that the CNN-based attendance system is a reliable and effective solution with the potential for widespread adoption, with future improvements aimed at further enhancing adaptability and reducing error rates.