Smart x-ray scanners using artificial neural networks

Abstract

Object detection and classification in Artificial Neural Networks (ANN) can play an important part in finding solutions for various tasks that are considered critical or time consuming if executed by humans. Also, object detection can be used in several applications and domains. One of which is physical security, where many illegal items, that should not pass through checkpoints, exist. In this sense, smart X-ray scanners can help in detecting any restricted object like weapons, knives or even drugs within a bag or held by an individual. This paper aims to provide a first step towards smart scanners where the MultiLayer Perceptron (MLP) is used to classify and detect two different types of illegal objects. Moreover, training of the network is performed using the Back Propagation algorithm, one of the most widely used algorithms in the MLP.