ADoCW: An Automated method for Detection of Concealed Weapon
Gaurav Raturi, Priya Rani, Sanjay Madan, Sonia Dosanjh · 2019
In technologically advanced era, surveillance is a proven method for the monitoring of the individual's activity in the crowd. Security of infrastructure, as well as individual, is one of the major concerns because of the influential growth of radical elements or suspicious persons in the society. Continuous manual monitoring of the CCTV surveillance is difficult and monotonous task, so there is an urgent requirement to develop an automated surveillance systems. The security surveillance system has potential to detect any kind of concealed object (like firearms or any weapon including knife, scissors etc.) which may pose a threat to the security. In this paper, we propose a novel framework for the detection and classification of concealed weapons through analysis of CCTV stream data. The classification framework is developed with the categorization of various concealed weapons through deep learning based object detection and classification techniques. For the detection of concealed weapon, multi-sensor stream data capturing framework is designed using sensor fusion techniques and also embedded with the feature extraction and segmentation of imsegmentation of images module. Faster R-CNN (Region-based Convolutional Neural Network) model is trained for classification ofages module. Faster R-CNN (Region-based Convolutional Neural Network) model is trained for classification of weapons over collected dataset. Finally, several directions of work and tasks are provided as future work for the various research communities.