Automated Weapon Detection Using Inception-v3: A CNN Approach

In today's world, the biggest threat to mankind is that of terrorism and governments all around the globe are facing hardships to cater to this menace. With the advancement of the internet and technology, there has been a rapid increase in the terrorist activities which includes cross border infiltration, bomb blasts, cyber-attacks and even public shooting. In public places, shooting has been identified as the biggest safety threat and can lead to catastrophic effects. In this paper, the focus is on automated detection in such situations. The proposed model uses Convolutional Neural Network (CNN) which is implemented with the concept of transfer learning. Inception-v3 Network has been used for image classification. CNN trains the model on given data sets so that the model learns about firearms and Inception-v3 Network helps in better training. Therefore, if any dangerous object gets detected in the closed circuit television systems (CCTV) footage, it notifies the person monitoring it. This will help governments save a lot of human effort and money.