SLX: Similarity Learning for X-Ray Screening and Robust Automated Disassembled Object Detection

Baggage screening is important in security-critical applications in airports for detecting threats, including firearms and parts of them. Existing approaches underperform to recognise prohibited objects that are disassembled, especially when learning from limited data and from images produced by different scanners with multi-view orientations. To address such limitations, in this paper, we develop the Similarity Learning X-ray screening (SLX) model for accurate and robust firearm component detection in cluttered scenes. We evaluate SLX on the X-ray Image Library (XIL) dataset that the UK Government has provided us with, for this research. SLX is based on a contrastive similarity learning approach combined with Out-of-Distribution (OoD) detection/ anomaly detection using a deep discriminative model, ResNet-152, for detecting and classifying forbidden items. The evaluation of SLX on the XIL dataset shows that it is effective, beneficial for detecting firearms and their parts, and outperforms other baseline models, on average, by approximately 12 points in accuracy.

I. Introduction

Baggage X-ray inspection is important for protecting public space from safety threatening, including terrorism. The screening of luggage is a core standard checking measure in airports [1], where security is of significant concern. Airports strive to automate detection to improve effectiveness and efficiency, even for firearm component detection within passengers' baggage, reducing errors and processing times. The problem is the following. From labelled data, we train a model that infers whether an image contains prohibited items, which are defined as items we would not want inside an airplane or parts of them, e.g. firearms and their components. The problem we consider in this paper is disassembled object detection in cluttered environments. The proposed discriminative Similarity Learning X-ray baggage screening (SLX) model aims at improving the accuracy for firearm and gun part detection.