Evaluating the Transferability and Adversarial Discrimination of Convolutional Neural Networks for Threat Object Detection and Classification within X-Ray Security Imagery

Gaus, Y.F.A.; Bhowmik, N.; Akcay, S.; Breckon, T.P.

Abstract

X-ray imagery security screening is essential to maintaining transport security against a varying profile of threat or prohibited items. Particular interest lies in the automatic detection and classification of weapons such as firearms and knives within complex and cluttered X-ray security imagery. Here, we address this problem by exploring various end-to-end object detection Convolutional Neural Network (CNN) architectures. We evaluate several leading variants spanning the Faster R-CNN, Mask R-CNN, and RetinaNet architectures to explore the transferability of such models between varying X-ray scanners with differing imaging geometries, image resolutions and material colour profiles. Whilst the limited availability of X-ray threat imagery can pose a challenge, we employ a transfer learning approach to evaluate whether such inter-scanner generalisation may exist over a multiple class detection problem. Overall, we achieve maximal detection performance using a Faster R-CNN architecture with a ResNet101 classification network, obtaining 0.88 and 0.86 of mean Average Precision (mAP) for a three-class and two class item from varying X-ray imaging sources. Our results exhibit a remarkable degree of generalisability in terms of cross-scanner performance (mAP: 0.87, firearm detection: 0.94 AP). In addition, we examine the inherent adversarial discriminative capability of such networks using a specifically generated adversarial dataset for firearms detection - with a variable low false positive, as low as 5%, this shows both the challenge and promise of such threat detection within X-ray security imagery.

Citation

Gaus, Y., Bhowmik, N., Akcay, S., & Breckon, T. (2019, December). Evaluating the Transferability and Adversarial Discrimination of Convolutional Neural Networks for Threat Object Detection and Classification within X-Ray Security Imagery. Presented at 18th IEEE International Conference on Machine Learning and Applications (ICMLA 2019), Boca Raton, Florida, USA

• Presentation Conference Type: Conference Paper (published)
• Conference Name: 18th IEEE International Conference on Machine Learning and Applications (ICMLA 2019)
• Start Date: Dec 16, 2019
• End Date: Dec 19, 2019
• Acceptance Date: Sep 21, 2019
• Online Publication Date: Feb 17, 2020
• Publication Date: 2019
• Deposit Date: Dec 20, 2019
• Publicly Available Date: Jun 4, 2020
• Publisher: Institute of Electrical and Electronics Engineers
• Pages: 420-425
• Book Title: 2019 18th IEEE International Conference On Machine Learning And Applications (ICMLA), Boca Raton, FL, USA, 2019.
• DOI: https://doi.org/10.1109/icmla.2019.00079
• Public URL: https://durham-repository.worktribe.com/output/1141343