Implementation of YOLO and Face Recognition on a Security Camera System for Visitor Detection in the Computer and Network Lab of UNS

Traditional security camera systems (CCTV cameras) only record video without automatic analysis, necessitating continuous human monitoring. This approach is inefficient and prone to human error. This research aims to implement the "You Only Look Once" (YOLO) method for object detection and face recognition in a security camera system to detect visitors to the computer and network laboratory at Sebelas Maret University (UNS). The study uses quantitative methods to measure the success rate of the system in identifying and tracking visitors entering the laboratory. The visitor counting program is implemented using Python, while the face recognition program utilizes API concepts in JavaScript to receive input from the visitor counting program. The results of the study show that the YOLO and face recognition programs perform well at a distance of 4 meters, outperforming other distance conditions. The programs work effectively under bright lighting conditions; however, under dark conditions, only the visitor counting program functions properly. Under optimal conditions, the object detection system achieves 99% accuracy, and the visitor counting system operates efficiently in both quiet (single person entry/exit) and busy (multiple people entry/exit) scenarios. The face recognition system also achieves 84% accuracy for 50 individuals across 5 classes under optimal conditions.