i-Guard: An AI-Powered Security System Leveraging YOLOv10

The need for automated detection of fire, bladed weapons, and firearms in high-security areas such as airports, police posts, offices, and industrial sites is critical for ensuring rapid and effective responses to potential threats. This paper introduces an AI-driven Intelligent Guard i.e., the i-Guard, a real-time security enhancement system leveraging YOLOv10 Nano to detect fire, bladed weapons (including knives, swords, machetes, and scissors), and firearms (such as handguns and rifles). To achieve this, a custom collected dataset of 20,000 images was used for model training, resulting in a mean average precision (mAP) of 66.7% with 1.1 ms of inference speed. The i-Guard system effectively identifies threats in real-time, providing a robust solution for improving safety and security in sensitive environments.