Skip to content

[IEEE GRSL] - DARecNet-BS: Unsupervised Dual Attention Reconstruction Network for Hyperspectral Band Selection

License

Notifications You must be signed in to change notification settings

ucalyptus/DARecNet-BS

Repository files navigation

DARecNet-BS

Pytorch implementation of our paper DARecNet-BS: Unsupervised Dual Attention Reconstruction Network for Hyperspectral Band Selection

Additional results are available in Supplementary

Usage - SSRN Classification

  • git clone https://github.com/ucalyptus/DARecNet-BS.git
  • cd DARecNet-BS/
  • conda create -n dabsenv python=3.6
  • conda activate dabsenv
  • pip install -r requirements.txt
  • cd SSRN/
  • python indian.py for Classification on Indian Pines Dataset.

See SSRN Settings for details on SSRN settings.

Usage - DARecNet-BS

Run the notebooks in the DARecNet-BS/ directory using https://colab.research.google.com

Architecture : DARecNet-BS

Citation

If you find useful DARecNet-BS code in your research, we would appreciate a citation to the original paper:

@article{roy2020bs,
	title={DARecNet-BS: Unsupervised Dual-Attention Reconstruction Network for Hyperspectral Band Selection},
	author={Roy, Swalpa Kumar and Das, Sayantan and Song, Tiecheng and Chanda, Bhabatosh},
	journal={IEEE Geoscience and Remote Sensing Letters},
	year={2020}
	}

Contact

Contact Sayantan Das for code issues or file an issue at this link

Acknowledgement

Part of this code is from a implementation of Classification of HSI using CNN by Zilong Zhong.

License

Copyright (c) 2020 Sayantan Das. Released under the MIT License. See LICENSE for details.

Releases

No releases published

Packages

No packages published