Implementation scripts of DSBMNet for hyperspectral anomaly detection. The manuscript is currently under peer review, and more details will be released soon.

Our experiments are implemented in:

• Python 3.12.9
• PyTorch 2.3.0
• torchvision 0.18.0
• numpy 1.26.0
• scipy 1.14.1

Put the dataset(.mat [data, map, E]) into ./dataset

Running main_HAD.py

-If you want to train and inference on your own dataset, add the dataset name and adjust parameters such as the learning rate and masked center block size.

The codes are based on UADNet and DIFF Transformer. Thanks for their awesome work.