• Afza fathima S
• Ms.Mathumitha.M
• Ishwarya.K
• Keerthika.E

Recent advancements in image recognition technology, particularly deep learning, have revolutionized security and home services by incorporating biometric data like fingerprints, iris scans, and face recognition. Among these, face recognition-based user authentication methods have gained significant traction. This study introduces a visitor authentication system using CCTV, Jetson Nano, and a webcam. In the preprocessing phase, CCTV captures face data with 7 key features for person identification. The collected dataset undergoes annotation and deep learning-based facial feature detection. If four or more features are detected, the data is classified as a person, and facial details are matched with stored user data using 81 feature vectors. Furthermore, security enhancements include logging visitor faces, visitor count, and visit times, bolstering the access control system. The paper implements this system on a Jetson Nano and evaluates its performance, focusing on accuracy and detection speed. The tiny-YOLOv3 model on Jetson Nano proves effective, achieving real-time face authentication with an average detection speed of 6.5 FPS and 86.3% accuracy. This study showcases a deep learning-based system for visitor authentication and access control, demonstrating its capabilities in real-time verification and user management.

Face recognition, Deep learning, Visitor authentication, Jetson nano

"CCTV-Powered Facial Authentication for Visitors", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.11, Issue 5, page no.a242-a253, May-2024, Available :http://www.jetir.org/papers/JETIR2405031.pdf

"CCTV-Powered Facial Authentication for Visitors", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.11, Issue 5, page no. ppa242-a253, May-2024, Available at : http://www.jetir.org/papers/JETIR2405031.pdf