• Rutik Nandkumar Ohal
• Vaishnavi More
• H. R. Kulkarni
• Priyanka Deshmukh

The integration of Artificial Intelligence (AI) and Computer Vision has led to significant advancements in biometric authentication systems. Among these, face recognition has become a prominent technology due to its contactless, non-intrusive, and highly accurate nature. This research presents a comprehensive AI-driven face recognition system designed for automated attendance management and secure authentication. The proposed framework utilizes deep learning architectures, including FaceNet, ArcFace, VGG-Face, and Hybrid CNN–Vision Transformer (ViT) models, to perform feature extraction, embedding generation, and identity matching with high discriminability. The system captures real-time video, detects and aligns faces, extracts embeddings, and compares them to a secured database using cosine similarity metrics. To enhance privacy, only embedding vectors are stored instead of raw facial images. The model demonstrates robust performance under varying conditions such as lighting, pose, and occlusion, achieving high accuracy while maintaining computational efficiency suitable for edge device deployment. Additionally, this research emphasizes ethical AI practices, including bias mitigation, data privacy, and compliance with emerging regulations like the EU AI Act. Overall, the system bridges the gap between theoretical research and real-world implementation, offering a scalable, secure, and ethical solution for institutions and enterprises aiming to automate authentication and attendance management.

Face Recognition, Artificial Intelligence, Deep Learning, Computer Vision, ArcFace, FaceNet, VGG-Face, Vision Transformer, Attendance Automation.

"Facial Recognition Using AI", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.13, Issue 1, page no.626-632, January-2026, Available :http://www.jetir.org/papers/JETIRHG06084.pdf

"Facial Recognition Using AI", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.13, Issue 1, page no. pp626-632, January-2026, Available at : http://www.jetir.org/papers/JETIRHG06084.pdf