The article discusses the development of a novel pipeline for automatic weapon detection using an ensemble of convolutional neural networks. This pipeline was created in response to the increasing prevalence of weapon misuse and gun violence. The paper highlights the need for an automatic system capable of detecting weapons in real-time from surveillance videos. This automated system is proposed as a solution to the tedious and time-consuming process of human monitoring. The research shows promising results, with the proposed pipeline producing an average increase of 5% in accuracy, specificity, and recall compared to state-of-the-art systems.
Publication date: 29 Sep 2023
Project Page: Not provided
Paper: https://arxiv.org/pdf/2309.16654
