A dataset proposed in "Asymmetric Contextual Modulation for Infrared Small Target Detection" https://arxiv.org/abs/2009.14530

SIRST is a dataset specially constructed for single-frame infrared small target detection, in which the images are selected from hundreds of infrared sequences for different scenarios.

The bounding box and semantic segmentation annotations are available now. The rest annotation forms will come soon.

The full learderboard will come soon. A comparison of 19 methods can be found at https://github.com/YimianDai/open-acm.

The full toolkit will come soon. Some implementations of our proposed methods can be found at DENTIST, ACM, and ALCNet.

Please cite our paper in your publications if our work helps your research. BibTeX reference is as follows.

@inproceedings{dai21acm,
  title   =  {Asymmetric Contextual Modulation for Infrared Small Target Detection},
  author  =  {Yimian Dai and Yiquan Wu and Fei Zhou and Kobus Barnard},
  booktitle =  {{IEEE} Winter Conference on Applications of Computer Vision, {WACV} 2021}
  year    =  {2021}
}