YOLO Fusion: Advanced Vision Model for Safety and Monitoring

YOLO Fusion: Advanced Vision Model for Safety and Monitoring proposes a unified real-time vision-based surveillance framework aimed at improving safety and automated monitoring across home, office, and public spaces. Conventional surveillance systems mainly act as passive recording devices and generally lack the intelligent analytical capabilities required for proactive threat detection and response. To overcome this limitation, the proposed approach utilizes the YOLOv8 deep learning architecture to combine multiple safety-related functions within a single integrated system.The framework enables real-time person detection, mask detection, recognition of family members, and identification of suspicious objects through a single-stage detection pipeline, which supports fast inference and efficient processing. To enhance the model’s reliability in practical scenarios, custom datasets were created and augmented to handle challenging real-world conditions. The system’s performance is assessed using widely accepted evaluation metrics, including accuracy, precision, recall, F1-score, and mean Average Precision (mAP). In addition, real-time performance indicators such as latency and frame rate are also considered. Experimental findings indicate strong overall system performance, achieving approximately 90% accuracy while maintaining stable real-time operation.