Local-Patch-Network-with-Global-Attention

Official PyTorch implementation of the paper entitled 'Local Patch Network with Global Attention for Infrared Small Target Detection'.

Network Architecture

Dataset

Two widely used public datasets can be downloaded from following links and we thank the authors for their nice works.

SIRST dataset: https://github.com/YimianDai/sirst

MFIRST dataset: https://github.com/wanghuanphd/MDvsFA_cGAN

Specifically, you need to organize the data folders & files as follows:

|
|--data/
|  |--MFIRST/
|  |  |--test/
|  |  |  |--00000.bmp
|  |  |  |--00001.bmp
|  |  |  |...
|  |  |  |--gt/
|  |  |  |  |--00000_gt.bmp
|  |  |  |  |--00001_gt.bmp
|  |  |  |  |...
|  |  |--train/
|  |  |  |--000000.bmp
|  |  |  |--000001.bmp
|  |  |  |...
|  |  |  |--gt/
|  |  |  |  |--000000_gt.bmp
|  |  |  |  |--000001_gt.bmp
|  |  |  |  |...
|  |--SIRST/
|  |  |--test/
|  |  |  |--Misc_6.bmp
|  |  |  |--Misc_8.bmp
|  |  |  |...
|  |  |  |--gt/
|  |  |  |  |--Misc_6_gt.bmp
|  |  |  |  |--Misc_8_gt.bmp
|  |  |  |  |...
|  |  |--train/
|  |  |  |--Misc_1.bmp
|  |  |  |--Misc_2.bmp
|  |  |  |...
|  |  |  |--gt/
|  |  |  |  |--Misc_1_gt.bmp
|  |  |  |  |--Misc_2_gt.bmp
|  |  |  |  |...

Please convert each mask to a binary image with only value 0 and 255, and name it as '[image name]_gt.bmp' before moving it into the 'gt' folder.

The division of training and testing set has been specified in the original dataset links and we just follow the authors' instruction.

Requirements

pip install -r requirements.txt

Run

python main.py

Citation

If you find this work useful for your research, please cite our paper:

@ARTICLE{9735292,
  author={Chen, Fang and Gao, Chenqiang and Liu, Fangcen and Zhao, Yue and Zhou, Yuxi and Meng, Deyu and Zuo, Wangmeng},
  journal={IEEE Transactions on Aerospace and Electronic Systems}, 
  title={Local Patch Network with Global Attention for Infrared Small Target Detection}, 
  year={2022},
  volume={},
  number={},
  pages={1-1},
  doi={10.1109/TAES.2022.3159308}}

Acknowledgements

This work is supported by the National Natural Science Foundation of China (No. 62176035, 61906025), the Natural Science Foundation of Chongqing, China (No. cstc2020jcyj-msxmX0835, cstc2021jcyj-bsh0155), the Macao Science and Technology Development Fund under Grant (061/2020/A2), the Science and Technology Research Program of Chongqing Municipal Education Commission under Grant (No. KJZD-K202100606, KJQN201900607, KJQN202000647, KJQN202100646), the China Postdoctoral Science Foundation (No. 2021MD703940).

Contact

If you have any technical questions, please contact:

Fang Chen

Email: fchen905@usc.edu or cfun.cqupt@outlook.com

License

This code is only freely available for non-commercial research use. If you have other purpose, please contact:

Chenqiang Gao

Email: gaochenqiang@gmail.com or gaocq@cqupt.edu.cn

Copyright: Chongqing University of Posts and Telecommunications

If you find some help for you, star is a good reward ^_^.