A PyTorch implementation of LiteSalNet for Remote Sensing Salient Object Detection.

TGRS has accepted this work!!!

We updated the LiteSalNet result graph！

We updated the LiteSalNet main model code！

The training code can be found at SeaNet. You only need to modify the model output. Dr.Li has done a very standard job in this regard.

🦉 Network Architecture

📝 Requirements

• Python 3.7
• PyTorch 1.9.0

🎉 Saliency maps

🏃‍♂️ Data

Download this dataset and put it into datasets.

LiteSalNet_data (code: AZXD)

🚀 Training

Run train_LiteSalNet.py.

🧩 Pre-trained model and testing

Download the following pre-trained model and put them in ./models/LiteSalNet/, then run test_LiteSalNet.py.

🎆 Result

Download Download the LiteSalNet model result graph (code:AZXD)

🛠️ Evaluation Tool

You can use the evaluation tool (MATLAB version) to evaluate the above saliency maps.

📖 Citation

    @ARTICLE{10945380,
            author = {Ai, Zhenxin and Luo, Huilan and Wang, Jianqin},
            title = {A Lightweight Multistream Framework for Salient Object Detection in Optical Remote Sensing},
            journal = {IEEE Transactions on Geoscience and Remote Sensing},
            volume = {63},
            year = {2025},
            doi = {10.1109/TGRS.2025.3555647},
            }

If you find this work useful for your research, feel free to cite it using the format above.