Contraband Materials Detection Within Volumetric 3D Computed Tomography Baggage Security Screening Imagery

TL;DR

This paper explores the application of 3D deep neural networks, including 3D U-Net and PointNet++, for detecting contraband materials in volumetric CT baggage scans, addressing a gap in 3D material detection performance.

Contribution

It introduces a novel approach by converting volumetric CT data into sparse point clouds to improve processing efficiency for material detection.

Findings

3D U-Net achieves high accuracy in material segmentation.

PointNet++ effectively processes sparse point cloud representations.

Both methods outperform traditional dense volume approaches.

Abstract

Automatic prohibited object detection within 2D/3D X-ray Computed Tomography (CT) has been studied in literature to enhance the aviation security screening at checkpoints. Deep Convolutional Neural Networks (CNN) have demonstrated superior performance in 2D X-ray imagery. However, there exists very limited proof of how deep neural networks perform in materials detection within volumetric 3D CT baggage screening imagery. We attempt to close this gap by applying Deep Neural Networks in 3D contraband substance detection based on their material signatures. Specifically, we formulate it as a 3D semantic segmentation problem to identify material types for all voxels based on which contraband materials can be detected. To this end, we firstly investigate 3D CNN based semantic segmentation algorithms such as 3D U-Net and its variants. In contrast to the original dense representation form of volumetric 3D CT data, we propose to convert the CT volumes into sparse point clouds which allows the use of point cloud processing approaches such as PointNet++ towards more efficient processing. Experimental results on a publicly available dataset (NEU ATR) demonstrate the effectiveness of both 3D U-Net and PointNet++ in materials detection in 3D CT imagery for baggage security screening.