Learning-Based Image Synthesis for Hazardous Object Detection in X-Ray Security Applications
X-ray baggage inspection has been widely used for maintaining airport and transportation security. Towards automated inspection, recent deep learning-based methods have attempted to detect hazardous objects directly from X-ray images. Since it is challenging to collect a large number of training ima...
Main Authors: | , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
IEEE
2021-01-01
|
Series: | IEEE Access |
Subjects: | |
Online Access: | https://ieeexplore.ieee.org/document/9552004/ |
_version_ | 1818724559941009408 |
---|---|
author | Hyo-Young Kim Sung-Jin Cho Seung-Jin Baek Seung-Won Jung Sung-Jea Ko |
author_facet | Hyo-Young Kim Sung-Jin Cho Seung-Jin Baek Seung-Won Jung Sung-Jea Ko |
author_sort | Hyo-Young Kim |
collection | DOAJ |
description | X-ray baggage inspection has been widely used for maintaining airport and transportation security. Towards automated inspection, recent deep learning-based methods have attempted to detect hazardous objects directly from X-ray images. Since it is challenging to collect a large number of training images from real-world environments, most previous learning-based methods rely on image synthesis for training data generation. However, these methods randomly combine foreground and background images, restricting the effectiveness of synthetic images for object detection. To solve this problem, in this paper, we propose a learning-based X-ray image synthesis method for object detection. Specifically, for each foreground object to be synthesized, we first estimate positions difficult to detect by the object detector. These positions and their corresponding confidence values are then used to construct a difficulty map, which is used for sampling the target foreground position for image synthesis. The performance analysis using various state-of-the-art object detectors shows that the proposed synthesis method can produce more useful training data compared with the conventional random synthesis method. |
first_indexed | 2024-12-17T21:28:21Z |
format | Article |
id | doaj.art-75acd3ce38d64cf392da9c4c4b87d193 |
institution | Directory Open Access Journal |
issn | 2169-3536 |
language | English |
last_indexed | 2024-12-17T21:28:21Z |
publishDate | 2021-01-01 |
publisher | IEEE |
record_format | Article |
series | IEEE Access |
spelling | doaj.art-75acd3ce38d64cf392da9c4c4b87d1932022-12-21T21:31:59ZengIEEEIEEE Access2169-35362021-01-01913525613526510.1109/ACCESS.2021.31162559552004Learning-Based Image Synthesis for Hazardous Object Detection in X-Ray Security ApplicationsHyo-Young Kim0https://orcid.org/0000-0001-8498-9650Sung-Jin Cho1https://orcid.org/0000-0001-9910-419XSeung-Jin Baek2https://orcid.org/0000-0003-0494-2372Seung-Won Jung3https://orcid.org/0000-0002-0319-4467Sung-Jea Ko4https://orcid.org/0000-0002-4875-7091School of Electrical Engineering, Korea University, Seoul, South KoreaSchool of Electrical Engineering, Korea University, Seoul, South KoreaSchool of Electrical Engineering, Korea University, Seoul, South KoreaSchool of Electrical Engineering, Korea University, Seoul, South KoreaSchool of Electrical Engineering, Korea University, Seoul, South KoreaX-ray baggage inspection has been widely used for maintaining airport and transportation security. Towards automated inspection, recent deep learning-based methods have attempted to detect hazardous objects directly from X-ray images. Since it is challenging to collect a large number of training images from real-world environments, most previous learning-based methods rely on image synthesis for training data generation. However, these methods randomly combine foreground and background images, restricting the effectiveness of synthetic images for object detection. To solve this problem, in this paper, we propose a learning-based X-ray image synthesis method for object detection. Specifically, for each foreground object to be synthesized, we first estimate positions difficult to detect by the object detector. These positions and their corresponding confidence values are then used to construct a difficulty map, which is used for sampling the target foreground position for image synthesis. The performance analysis using various state-of-the-art object detectors shows that the proposed synthesis method can produce more useful training data compared with the conventional random synthesis method.https://ieeexplore.ieee.org/document/9552004/Deep learningneural networkobject detectionX-rayinspection |
spellingShingle | Hyo-Young Kim Sung-Jin Cho Seung-Jin Baek Seung-Won Jung Sung-Jea Ko Learning-Based Image Synthesis for Hazardous Object Detection in X-Ray Security Applications IEEE Access Deep learning neural network object detection X-ray inspection |
title | Learning-Based Image Synthesis for Hazardous Object Detection in X-Ray Security Applications |
title_full | Learning-Based Image Synthesis for Hazardous Object Detection in X-Ray Security Applications |
title_fullStr | Learning-Based Image Synthesis for Hazardous Object Detection in X-Ray Security Applications |
title_full_unstemmed | Learning-Based Image Synthesis for Hazardous Object Detection in X-Ray Security Applications |
title_short | Learning-Based Image Synthesis for Hazardous Object Detection in X-Ray Security Applications |
title_sort | learning based image synthesis for hazardous object detection in x ray security applications |
topic | Deep learning neural network object detection X-ray inspection |
url | https://ieeexplore.ieee.org/document/9552004/ |
work_keys_str_mv | AT hyoyoungkim learningbasedimagesynthesisforhazardousobjectdetectioninxraysecurityapplications AT sungjincho learningbasedimagesynthesisforhazardousobjectdetectioninxraysecurityapplications AT seungjinbaek learningbasedimagesynthesisforhazardousobjectdetectioninxraysecurityapplications AT seungwonjung learningbasedimagesynthesisforhazardousobjectdetectioninxraysecurityapplications AT sungjeako learningbasedimagesynthesisforhazardousobjectdetectioninxraysecurityapplications |