1. Title

ATIS-Driven 3DCNet: A Novel Three-Stream Hyperspectral Fusion Framework with Knowledge from Downstream Classification Performance. Remote Sensing, 2025.

2. Network Architecture

3. Please cite

If you use any part of this code, please cite the following paper:

@article{zhang2025atis,
  title={ATIS-Driven 3DCNet: A Novel Three-Stream Hyperspectral Fusion Framework with Knowledge from Downstream Classification Performance},
  author={Zhang, Quan and Long, Jian and Li, Jun and Li, Chunchao and Si, Jianxin and Peng, Yuanxi},
  journal={Remote Sensing},
  volume={17},
  number={5},
  pages={825},
  year={2025},
  publisher={Multidisciplinary Digital Publishing Institute}
}

4. Datasets

https://github.com/ZhangQuan-hub/HyperspectralDatasets.git

5. Settings

• Python 3.11

• Pytorch 2.1.0

• CUDA 12.1

Note: Version inconsistency does not have a significant impact on performance.

6. Auxiliary code for Training and Acknowledgment

This repo is based on SSRNET, which is an exellent work. All training architectures are based on SSRNET. You should only change the model.py file to our proposed model.py.

7. Contact

If you have any questions, please contact me at zq@nudt.edu.cn