Firearm Detection and Segmentation using an Ensemble of Semantic Neural Networks

No ratings

Presented at EISIC 2019 by

Alexander Egiazarov

Vasileios Mavroeidis

Fabio Vishi

In recent years we have seen an upsurge in terror attacks around the world. Such attacks usually happen in publicplaces with large crowds to cause the most damage possible and get the most attention. Even though surveillancecameras are assumed to be a powerful tool, their effect in preventing crime is far from clear due to either limitationin the ability of humans to vigilantly monitor video surveillance or for the simple reason that they are operatingpassively. In this paper, we present a weapon detection system based on an ensemble of semantic ConvolutionalNeural Networks that decomposes the problem of detecting and locating a weapon into a set of smaller problemsconcerned with the individual component parts of a weapon. This approach has computational and practicaladvantages: a set of simpler neural networks dedicated to specific tasks requires less computational resources andcan be trained in parallel; the overall output of the system given by the aggregation of the outputs of individualnetworks can be tuned by a user to trade-off false positives and false negatives; finally, according to ensembletheory, the output of the overall system will be robust and reliable even in the presence of weak individual models.We evaluated our system running simulations aimed at assessing the accuracy of individual networks and the wholesystem. The results on synthetic data and real-world data are promising, and they suggest that our approach mayhave advantages compared to the monolithic approach based on a single deep convolutional neural network.