• 2026-05-08
• J.O. Adeogo, O.J. Olaluyi, J.O. Okunlola, K.O. Olusuyi, O.F. Adeoye
• Volume 2, Issue 1

Abstract

In this study, the implementation of an automated biometric facial identification for student examination permit designed to replace traditional manual and biometric attendance methods with a more secure and efficient solution. The system integrates Python-based computer vision techniques, combining OpenCV for image processing with a deep-learning enabled face recognition model for accurate identity verification. Instead of depending solely on conventional biometric devices prone to impersonation and physical wear, the proposed system uses real-time webcam input to detect, encode, and match employee facial features against a trained dataset. For performance evaluation, we compare recognition accuracy across multiple environmental conditions, including variations in lighting, facial occlusion, and camera resolution. We also analyze the system’s behavior relative to existing attendance methods in terms of reliability, processing speed, and data integrity. Experimental results demonstrate that the Python-based system delivers high accuracy under optimal lighting and maintains robust performance in most workplace scenarios, outperforming manual attendance mechanisms and offering significant improvements in transparency and security. Although recognition accuracy reduces under extremely poor illumination or heavy occlusion, the system still maintains acceptable performance suitable for real-world deployment.