Real-Time Threat Detection from Surveillance Cameras using Machine Learning
Researchers have developed a real-time threat detection system for surveillance cameras, utilizing machine learning to identify weapons like guns, knives, and blunt objects. The system was trained on a combined dataset of nearly 8,000 images, including a custom collection of blunt objects. A YOLOv8 model was employed for object detection, demonstrating improved accuracy and recall for blunt objects with extended training. AI
IMPACT This research could enhance public safety by enabling automated threat detection in surveillance systems.