An approach for adaptive automatic threat recognition within 3D computed tomography images for baggage security screening

Abstract

BACKGROUND: The screening of baggage using X-ray scanners is now routine in aviation security with automatic threat detection approaches, based on 3D X-ray computed tomography (CT) images, known as Automatic Threat Recognition (ATR) within the aviation security industry. These current strategies use pre-defined threat material signatures in contrast to adaptability towards new and emerging threat signatures. To address this issue, the concept of adaptive automatic threat recognition (AATR) was proposed in previous work.OBJECTIVE: In this paper, we present a solution to AATR based on such X-ray CT baggage scan imagery. This aims to address the issues of rapidly evolving threat signatures within the screening requirements. Ideally, the detection algorithms deployed within the security scanners should be readily adaptable to different situations with varying requirements of threat characteristics (e.g., threat material, physical properties of objects).METHODS: We tackle this issue using a novel adaptive machine learning methodology with our solution consisting of a multi-scale 3D CT image segmentation algorithm, a multiclass support vector machine (SVM) classifier for object material recognition and a strategy to enable the adaptability of our approach. Experiments are conducted on both open and sequestered 3D CT baggage image datasets specifically collected for the AATR study.RESULTS: Our proposed approach performs well on both recognition and adaptation. Overall our approach can achieve the probability of detection around 90% with a probability of false alarm below 20%. Figure 1: Threat objects in 2D X-ray images [2]. Left: threat object signatures of specific appearance (e.g., firearms) are easy to identify. Right: threat materials are more challenging to recognize in cluttered X-ray images.

On the Evaluation of Prohibited Item Classification and Detection in Volumetric 3D Computed Tomography Baggage Security Screening Imagery

“…Following previous works in [5], we report the results of PD and PFA in Table III. By comparing with the results of [5], both the point cloud based methods and 3D U-Net based methods can achieve comparable or better performance. Although PointNet gives worse results than the traditional method, its variant PointNet++ can achieve comparable PD of 92% and PFA of 24%.…”

Volatile Liquid Detection by Terahertz Technologies

“…X-ray imaging-based apparatus provides a major tool in checked baggage inspection, as it can detect the form and density of items within luggage as well as other material dependent parameters. Recently, conventional medical computed tomography (CT) scans have been developed and introduced to some European airports, where 3D images of items are computing-processed to combine hundreds of individual X-ray measurements from different angles [2,3]. The technology offers more detailed and comprehensive image quality; however, it was not able to highlight whether a substance was a solid or a liquid.…”

On the Evaluation of Semi-Supervised 2D Segmentation for Volumetric 3D Computed Tomography Baggage Security Screening

“…Automatic threat detection in baggage security screening enables more efficient and safer transportation. Recent advances in deep learning and image processing make it possible for automatic threat detection in 2D X-ray and 3D Computed Tomography (CT) baggage security imagery [1]- [6]. Stateof-the-art deep learning models for image classification and object detection trained on ImageNet [7] and MS-COCO [8] can be transferred to X-ray imagery by fine-tuning the pretrained models on relatively smaller X-ray datasets.…”

“…Efforts were made towards material classification within 3D CT imagery in [6] and [14] using a hand-engineered framework and 3D CNN based deep learning techniques, respectively. One limitation of these existing works is the expensive computational cost introduced by the use of 3D CNN.…”