SSUMamba

Code of "SSUMamba: Spatial-Spectral Selective State Space Model for Hyperspectral Image Denoising"

This code is strongly based on T3SC (https://github.com/inria-thoth/T3SC)

The data can be download at https://njusteducn-my.sharepoint.com/:f:/g/personal/321106020189_njust_edu_cn/ErC7SGyqgdlEm86mrlbZ-EwBUSPG9UJchCT-997L0IiNMA?e=kjfM6j .

After you download the datasets, please modify the dir path in mamba/config/test/*.yaml

• Please install the requriment in ./ssumamba.yaml.

• Install PyTorch 2.1.1+cu118

  pip install torch==2.1.1 torchvision==0.16.1 torchaudio==2.1.1 --index-url https://download.pytorch.org/whl/cu118

• Install causal_conv1d and mamba

  pip install -e requirements/causal-conv1d
  pip install -e requirements/mamba

Then, you can try train*_sscs.sh or test*_sscs.sh to evaluate the SSUMamba with spatial-spectral continuous scan in our paper.