Real-Time Threat Detection from Surveillance Cameras using Machine Learning
2026Gajendra Mandal, J. P. Patra, Priyansh Mahant
RECCAP 2026
This paper presents a real-time surveillance framework for detecting weapons and blunt objects using a YOLOv8-based object detection model. The proposed system combines a custom Indian blunt-object dataset with a publicly available gun and knife dataset to create a dataset of 7,959 images. Experimental results demonstrate improved detection performance and real-time inference, making the system suitable for deployment in campuses, public spaces, and transportation environments.
computer visionYOLOv8object detectionsurveillanceweapon detectionmachine learningdeep learningsecurity
View Paper