Awesome

TransENet for remote sensing image super-resolution

Official Pytorch implementation of the paper "Transformer-based Multi-Stage Enhancement for Remote Sensing Image Super-Resolution" accepted by IEEE TGRS.

Convolutional neural networks have made great breakthrough in recent remote sensing image super-resolution tasks. Most of these methods adopt upsampling layers at the end of the models to perform enlargement, which ignores feature extraction in the high-dimension space and thus limits super-resolution performance. To address this problem, we propose a new super-resolution framework for remote sensing image to enhance the high-dimensional feature representation after the upsampling layers. We name the proposed method as Transformer-based Enhancement Network (TransENet), where transformers are introduced to exploit features at different levels. The core of the TransENet is a transformer-based multi-stage enhancement structure which can be combined with traditional super-resolution frameworks to fuse multi-scale high/low-dimension features. Specifically, in this structure, the encoders aim to embed the multi-level features in the feature extraction part and the decoders are used to fuse these encoded embeddings. Experimental results demonstrate that our proposed TransENet can improve super-resolved results and obtain superior performance over several state-of-the-art methods.

Requirements

• Python 3.6+
• Pytorch>=1.6
• torchvision>=0.7.0
• einops
• matplotlib
• cv2
• scipy
• tqdm
• scikit

Installation

Clone or download this code and install aforementioned requirements

cd codes

Train

Download the UCMerced dataset[Baidu Drive,password:terr][Google Drive]and AID dataset[Baidu Drive,password:id1n][Google Drive], they have been split them into train/val/test data, where the original images would be taken as the HR references and the corresponding LR images are generated by bicubic down-sample.

# x4
python demo_train.py --model=TRANSENET --dataset=UCMerced --scale=4 --patch_size=192 --ext=img --save=TRANSENETx4_UCMerced
# x3
python demo_train.py --model=TRANSENET --dataset=UCMerced --scale=3 --patch_size=144 --ext=img --save=TRANSENETx3_UCMerced
# x2
python demo_train.py --model=TRANSENET --dataset=UCMerced --scale=2 --patch_size=96 --ext=img --save=TRANSENETx2_UCMerced

The train/val data pathes are set in data/init.py

Test

The trained TransENet model on UCMerced and AID datasets can be found here [Baidu Drive, password:w7ct][Google Drive]. The test data path and the save path can be edited in demo_deploy.py

# x4
python demo_deploy.py --model=TRANSENET --scale=4
# x3
python demo_deploy.py --model=TRANSENET --scale=3
# x2
python demo_deploy.py --model=TRANSENET --scale=2

Evaluation

Compute the evaluated results in term of PSNR and SSIM, where the SR/HR paths can be edited in calculate_PSNR_SSIM.py

cd metric_scripts 
python calculate_PSNR_SSIM.py

Citation

If you find this code useful for your research, please cite our paper:

@article{lei2021transformer,
  title={Transformer-based Multi-Stage Enhancement for Remote Sensing Image Super-Resolution},
  author={Lei, Sen and Shi, Zhenwei and Mo, Wenjing},
  journal={IEEE Transactions on Geoscience and Remote Sensing},
  year={2021},
  publisher={IEEE}
}

Acknowledgements

This code is built on RCAN (Pytorch) and EDSR (Pytorch). We thank the authors for sharing the codes.