A New Approach for Baggage Inspection by using Deep Convolutional Neural Networks

In recent years, the use of x-ray equipment in different security point has increased. This equipment is heavily used at airports to control the baggage and bag of peoples. With this control, criminals are detected and terrorist acts can be prevented. This task is done by security officers at security points. It requires high concentration for the detection of threat objects. The manual operation of this process is both tedious and requires constant attention. There are many problems in the control with computer based automated systems. Because the position of the object in the baggage, overlapping with other objects makes the checking process difficult. In this study, a deep learning-based method for baggage control system was proposed by using x-ray images. The proposed method uses regions with convolutional neural networks for threat object detection. First, each objects in the images are labelled. Afterwards, the location of the image and bounding boxes of objects are given to regions with convolutional neural networks. The threat objects are detected and recognized in the last step. The proposed method is tested on a baggage inspection dataset and satisfied results are obtained.