OPIXray dataset and DOAM-O Module

This repository is OPIXray dataset and the official implementation of De-occlusion Attention Module(DOAM-O).

Paper link: https://arxiv.org/abs/2004.08656.

Dataset Download link:

(China mainland, BaiduNetdisk) https://pan.baidu.com/s/18LX4C_ZfJUWMF19L0fNdBA password: i17p

(China mainland, BaiduNetdisk, Background Images) https://pan.baidu.com/s/1mc_XNEJlsF_0Rhi8-OHfbA password: bnqz

(Other area, Google Drive) https://drive.google.com/file/d/1lqgH4am3f30tv2FHZrHQNUCVyORB3PLb/view?usp=sharing

V2 (submitted to IEEE T-NNLS)

OPIXray dataset:

DOAM-O framework:

V1 (published in ACM Multimedia 2020)

DOAM model:

Table of Contents

• Overview
• Dataset
• Install
• Usage
• Results
• Contributing
• Acknowledgement
• License

Overview

Security inspection often deals with a piece of baggage or suitcase where objects are heavily overlapped with each other, resulting in an unsatisfactory performance for prohibited items detection in X-ray images.

In this work, first, we contribute a high-quality dataset named OPIXray, each of which is annotated manually by professional inspectors from an international airport. To the best of our knowledge, this is the first dataset specifically designed for object detection in security inspection. Second, we propose a De-occlusion Attention Module (DOAM) that can be inserted as a plug-and-play module into most detectors, aiming at detecting occluded prohibited items in X-ray images. Finally, we evaluate our method on the OPIXray dataset and compare it to several baselines, including popular methods for detection and attention mechanisms. As is shown from the results, our proposed method significantly outperforms existing models.

Dataset

V1

In this project, we build the first dataset specially designed for occluded prohibited items detection in security inspection. Each picture was reshape to 300 * 300 during training.

• All images of OPIXray dataset are scanned by security inspection machine and annotated manually by professional inspectors from an international airport, and the standard of annotating is based on the standard of training security inspectors.

• OPIXray dataset contains a total of 8885 X-ray images(7019 for training, 1776 for testing), including 5 categories of cutters, namely, Folding Knife, Straight Knife, Scissor, Utility Knife, Multi-tool Knife.

  

• In order to study the impact brought by object occlusion levels, we divide the testing set into three subsets and name them Occlusion Level 1 (OL1), Occlusion Level 2 (OL2) and Occlusion Level 3 (OL3), where the number indicates occlusion level of prohibited items in images.

  

• The information structure of annotation file is as follows: image name, category, top-left position of prohibited item (x1, y1), bottom-right position of prohibited item (x2, y2).

V2

• On the basis of the original OPIXray dataset, we added 10000 background pictures excluding the above dangerous items.

  

Install

1. Install pytorch

   The code is tested on python3.5 and pytorch1.3.1

2. Clone the repository

   git clone https://github.com/OPIXray-author/OPIXray.git
   

3. Access OPIXray dataset

   OPIXray dataset is available only for ACADEMIC PURPOSE, if you are interested in or have any questions about our dataset, feel free to contact [email protected]. (Both V1 and V2 datasets are available)

Usage

Training

1. change root to DOAM

   cd DOAM
   

2. modify OPIXray_ROOT in data/OPIXray.py :

   OPIXray_ROOT = 'OPIXray_Dataset/train/'	# path to trainset of OPIXray Dataset
   

3. run train.py

   python train.py --image_sets 'OPIXray_Dataset/train/train_knife.txt' --transfer './weights/ssd300_mAP_77.43_v2.pth' --save_folder './checkpoints'
   

   • --save_folder: the path to save the checkpoint while training.
   • --dataset_root: the path to the training dataset.
   • --image_sets: the path to a TXT file that saves all the picture names used for training.
   • --transfer: indicates the path to SSD pre-trained weight(available at here).

   Notes: The model can converge after 60 ~ 80 epoch training, and long-time training will lead to over fitting.
   

Testing

1. change root to DOAM

   cd DOAM
   

2. modify OPIXray_ROOT in data/OPIXray.py :

   OPIXray_ROOT = "OPIXray_Dataset/test/"	# path to testset of OPIXray Dataset
   

3. run test.py

   python test.py --imagesetfile 'OPIXray_Dataset/test/test_knife.txt' --trained_model './weights/DOAM.pth' 
   

   • --imagesetfile: the path to a TXT file that saves all the picture names used for testing.
     • test_knife.txt: total testset
     • test_knife-1.txt: occlusion level 1 (no or slight occlusion).
     • test_knife-2.txt: occlusion level 2 (partial occlusion).
     • test_knife-3.txt: occlusion level 3 (severe or full occlusion).
   • our model is available at here, password is 2abm.

License

The images and the corresponding annotations in OPIXray Dataset can be used ONLY for academic purposes, NOT for commercial purposes.

Citation

If this work helps your research, please cite the following paper.

@inproceedings{wei2020occluded,
  title={Occluded prohibited items detection: An x-ray security inspection benchmark and de-occlusion attention module},
  author={Wei, Yanlu and Tao, Renshuai and Wu, Zhangjie and Ma, Yuqing and Zhang, Libo and Liu, Xianglong},
  booktitle={Proceedings of the 28th ACM International Conference on Multimedia},
  pages={138--146},
  year={2020}
}

Copyright © 2020 by State Key Lab of Software Development Environment, Beihang University

All rights reserved.