Toward Automatic Threat Recognition for Airport X-ray Baggage Screening\n with Deep Convolutional Object Detection

Abstract

For the safety of the traveling public, the Transportation Security\nAdministration (TSA) operates security checkpoints at airports in the United\nStates, seeking to keep dangerous items off airplanes. At these checkpoints,\nthe TSA employs a fleet of X-ray scanners, such as the Rapiscan 620DV, so\nTransportation Security Officers (TSOs) can inspect the contents of carry-on\npossessions. However, identifying and locating all potential threats can be a\nchallenging task. As a result, the TSA has taken a recent interest in deep\nlearning-based automated detection algorithms that can assist TSOs. In a\ncollaboration funded by the TSA, we collected a sizable new dataset of X-ray\nscans with a diverse set of threats in a wide array of contexts, trained\nseveral deep convolutional object detection models, and integrated such models\ninto the Rapiscan 620DV, resulting in functional prototypes capable of\noperating in real time. We show performance of our models on held-out\nevaluation sets, analyze several design parameters, and demonstrate the\npotential of such systems for automated detection of threats that can be found\nin airports.\n