Skip to content

Latest commit

 

History

History
58 lines (50 loc) · 1.85 KB

README.md

File metadata and controls

58 lines (50 loc) · 1.85 KB

Spectrum-oriented Point-supervised Saliency Detector for Hyperspectral Images [TIM 2024]

by Peifu Liu, Tingfa Xu, Guokai Shi, Jingxuan Xu, Huan Chen, and Jianan Li.

Requirements

conda create --name SPSD --file requirements.txt

pydensecrf should be installed by:

pip install cython
conda install -c conda-forge pydensecrf

Getting Started

Prepare Data

Please download from Baidu Netdisk and place them in the Data folder. The folder structure should be as follows:

/Data
    /color
        /train
        /test
    /GT
        /train
        /test
    /hyperspectral
        /train
        /test
    /spec_sal
        /train
        /test
    /training
        /edge_gt
        /mask
        /pseudo-label
    train.txt
    test.txt

Note: The hyperspectral data we use is obtained by downsampling the raw data. For the original data, please refer to HSOD-BIT.

Training, Testing, and Evaluation

Just run the following command:

bash run.sh

We also provide our trained model and detection results. Please download from Baidu Netdisk for reproduction.

Acknowledgement

We refer to the following repositories:

Thanks for their great work!

License

This project is licensed under the LICENSE.md.