{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/unexpected-item-in-the-bagging-area-anomaly","title":"‘Unexpected item in the bagging area’: Anomaly Detection in X-ray Security Images","arxiv_id":null,"date":"2018-11-16","proceeding":"IEEE TRANSACTIONS ON INFORMATION FORENSICS AND SECURITY 2018 11","authors":["Lewis D. Griffin","Matthew Caldwell","Jerone T. A. Andrews","Helene Bohler"],"abstract":"The role of Anomaly Detection in X-ray security\r\nimaging, as a supplement to targeted threat detection, is described;\r\nand a taxonomy of anomalies types in this domain is presented.\r\nAlgorithms are described for detecting appearance anomalies, of\r\nshape, texture and density; and semantic anomalies of object\r\ncategory presence. The anomalies are detected on the basis of\r\nrepresentations extracted from a convolutional neural network\r\npre-trained to identify object categories in photographs: from the\r\nfinal pooling layer for appearance anomalies, and from the logit\r\nlayer for semantic anomalies. The distribution of representations\r\nin normal data are modelled using high-dimensional, full-\r\ncovariance, Gaussians; and anomalies are scored according to\r\ntheir likelihood relative to those models. The algorithms are tested\r\non X-ray parcel images using stream-of-commerce data as the\r\nnormal class, and parcels with firearms present as examples of\r\nanomalies to be detected. Despite the representations being learnt\r\nfor photographic images, and the varied contents of stream-of-\r\ncommerce parcels; the system, trained on stream-of-commerce\r\nimages only, is able to detect 90% of firearms as anomalies, while\r\nraising false alarms on 18% of stream-of-commerce.","url_abs":"https://discovery.ucl.ac.uk/id/eprint/10062484/1/2nd%20revision%20in%20template%2002.pdf","url_pdf":"https://discovery.ucl.ac.uk/id/eprint/10062484/1/2nd%20revision%20in%20template%2002.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[],"tasks":[{"task_slug":"anomaly-detection","task_name":"Anomaly Detection"}],"methods":[],"datasets_introduced":[{"slug":"compass-xp","name":"COMPASS-XP","full_name":""}],"methods_introduced":[],"results":[],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}