1 May 2017 Representation-learning for anomaly detection in complex x-ray cargo imagery
Author Affiliations +
Existing approaches to automated security image analysis focus on the detection of particular classes of threat. However, this mode of inspection is ineffectual when dealing with mature classes of threat, for which adversaries have refined effective concealment techniques. Furthermore, these methods may be unable to detect potential threats that have never been seen before. Therefore, in this paper, we investigate an anomaly detection framework, at X-ray image patch-level, based on: (i) image representations, and (ii) the detection of anomalies relative to those representations. We present encouraging preliminary results, using representations learnt using convolutional neural networks, as well as several contributions to a general-purpose anomaly detection algorithm based on decision-tree learning.
Conference Presentation
© (2017) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Jerone T. A. Andrews, Nicolas Jaccard, Thomas W. Rogers, Lewis D. Griffin, "Representation-learning for anomaly detection in complex x-ray cargo imagery", Proc. SPIE 10187, Anomaly Detection and Imaging with X-Rays (ADIX) II, 101870E (1 May 2017); doi: 10.1117/12.2261101; https://doi.org/10.1117/12.2261101


Hybrid object detection system for x-ray radiographs
Proceedings of SPIE (October 03 2016)
Additive Dirichlet models for projectional images
Proceedings of SPIE (February 24 2012)
Fuzzy iterative image segmentation with recursive merging
Proceedings of SPIE (November 01 1992)
X-ray backscatter imaging
Proceedings of SPIE (April 25 2008)

Back to Top