INVESTIGATING AN EXCEPTIONAL LEAP IN SURVEILLANCE TECHNOLOGY USING YOLO V8 ALGORITHM FOR DETECTING AND PROCESSING VISUAL IMAGES OF DIFFERENT OBJECTS

This study presents a comprehensive evaluation of the YOLO V8 framework within the context of surveillance technology, focusing on its performance in three critical use cases: license plate detection, face recognition, and suspicious activity detection. Our proposed model utilizes camera-captured input data processed through the YOLO V8 architecture. Experiments benchmarked YOLO V8 against YOLO V7, Faster R-CNN, and SSD using publicly available datasets: OpenALPR Benchmark Dataset for license plates, Labeled Faces in the Wild (LFW) for face recognition, and UCF-Crime Dataset for suspicious activity detection. Advanced frameworks like TensorFlow and PyTorch were employed, along with cutting-edge GPU architectures to optimize training and inference speeds. Performance was rigorously evaluated based on key metrics including mean Average Precision (mAP), precision, recall, F1 score, and inference time. Results demonstrated that YOLO V8 outperformed competing models across all metrics, highlighting its effectiveness for real-time detection in surveillance applications.