YOLO-based threat object detection in x-ray images

• Main Authors: Galvez, Reagan L., Dadios, Elmer P., Bandala, Argel A., Vicerra, Ryan Rhay P.
• Format: text
• Published: Animo Repository 2019
• Subjects: X-rays; Neural networks (Computer science); Image converters; Manufacturing
• Online Access: https://animorepository.dlsu.edu.ph/faculty_research/2049
• Institution: De La Salle University

Manual detection of threat objects in an X-ray machine is a tedious task for the baggage inspectors in airports, train stations, and establishments. Objects inside the baggage seen by the X-ray machine are commonly occluded and difficult to recognize when rotated. Because of this, there is a high chance of missed detection, particularly during rush hour. As a solution, this paper presents a You Only Look Once (YOLO)based object detector for the automated detection of threat objects in an X-ray image. The study compared the performance between using transfer learning and training from scratch in an IEDXray dataset which composed of scanned Xray images of improvised explosive device (IED) replicas. The results of this research indicate that training YOLO from scratch beats transfer learning in quick detection of threat objects. Training from scratch achieved a mean average precision (mAP) of 45.89% in 416×416 image, 51.48% in 608×608 image, and 52.40% in a multi-scale image. On the other hand, using transfer learning achieved only an mAP of 29.54% while 29.17% mAP in a multi-scale image. © 2019 IEEE.