AI-Based Weapon Detection System for the Prevention of Any Potential Crime
Husna Tabassum, Sandeep Kumar Singh, Sami Ibrahim, Syeda Sama Hussain, Heena Firdaus · Zenodo (CERN European Organization for Nuclear Research) · 2023
Security cameras and video surveillance systems have become important infrastructures to ensure the safety and security of the general public. Due to the growing demand for safety and security, there is a need for video surveillance systems that can recognize and interpret the scene and send in the required response for the same. Manual discovery of dangerous things is wearying, as a result, this idea was proposed. In this, we detect any harmful weapon and send an alert to the respective authority so that they can take the appropriate action. This work aims to develop a low-cost, efficient, and artificial intelligence-based solution for the real-time detection and recognition of weapons in surveillance videos under different scenarios. The system was developed based on TensorFlow and is preliminarily tested. This implementation makes use of two types of datasets. One dataset has pre-labelled images and the other one is a set of images, which are labelled manually. This has versatile applications worldwide.