X-ray Baggage Screening and Artificial Intelligence (AI)
A technical review of machine learning techniques for X-ray baggage screening.
The aim of this report is to review the scientific literature and state of the art regarding the application of machine learning techniques in X-ray security screening of baggage. We begin by reviewing X-ray baggage screening technology, followed by a discussion on the importance of human-machine interaction. The different approaches to measuring and describing the performance of AI algorithms are summarised, and an overview of existing databases of X-ray images is given. An overview of image enhancement and threat detection using classical machine learning techniques is described followed by data augmentation then deep learning. We also describe some applications of machine learning to materials classification (as opposed to object detection). The report concludes by discussing some horizontal issues concerning the application of AI in X-ray baggage screening, including testing of algorithm performance, the need for large, harmonised databases of images, data scarcity, transparency, and explainability.
VUKADINOVIC Danijela;
ANDERSON David;
2022-06-15
Publications Office of the European Union
JRC129088
978-92-76-53494-5 (online),
1831-9424 (online),
EUR 31123 EN,
OP KJ-NA-31123-EN-N (online),
https://publications.jrc.ec.europa.eu/repository/handle/JRC129088,
10.2760/46363 (online),
Additional supporting files
| File name | Description | File type | |
|---|
