TSTTD

Efficient replication for the paper: "Triplet Spectralwise Transformer Network for Hyperspectral Target Detection", TGRS, 2023.

Dependencies

• Python 3.11.0
• PyTorch 2.0
• NVIDIA TITAN RTX

Demo

Demo for TSTTD

1. You can directly test the model on San Diego.

2. If you want to train it on other datasets, please change the state to "train" and modify the data path and number of bands.

3. After finishing training, you can change the state to "eval" to test.

Citations

This code is replicated only for academic use.

If you find the code helpful in your research or work, please cite the original paper:

@ARTICLE{10223236,
  author={Jiao, Jinyue and Gong, Zhiqiang and Zhong, Ping},
  journal={IEEE Transactions on Geoscience and Remote Sensing}, 
  title={Triplet Spectralwise Transformer Network for Hyperspectral Target Detection}, 
  year={2023},
  volume={61},
  number={},
  pages={1-17},
  keywords={Training;Transformers;Feature extraction;Hyperspectral imaging;Detectors;Object detection;Task analysis;Balanced learning;hyperspectral image;spectralwise transformer;target detection;triplet network},
  doi={10.1109/TGRS.2023.3306084}}

Replication Information:

Dunbin Shen,

Dalian University of Technology,

May 5, 2024