This repo is the official implementation of 'Decomposition-based Unsupervised Domain Adaptation for Remote Sensing Image Semantic Segmentation'.
Please download the official pre-trained weights from 'Swin Transformer'--'ImageNet-1K and ImageNet-22K Pretrained Swin-V2 Models'--'SwinV2-B*' and 'SwinV2-L*'. The last cloumn '1K model' is what we used.
Run the GLGAN by: python GLGAN_*.py
Run the decomposition-based GLGAN by: python FDGLGAN_*.py
Please cite our paper if you find it is useful for your research.
@article{ma2024frequency,
author={Ma, Xianping and Zhang, Xiaokang and Ding, Xingchen and Pun, Man-On and Ma, Siwei},
journal={IEEE Transactions on Geoscience and Remote Sensing},
title={Decomposition-Based Unsupervised Domain Adaptation for Remote Sensing Image Semantic Segmentation},
year={2024},
volume={62},
number={},
pages={1-18}
}