KO
|
EN
gitlite — search
Search
#python
#java
#python3
#arduino
#golang
#machine-learning
#rust
#html
#flask
#javascript
#seismology
#nodejs
CTNet
★ 15
Open GitHub ↗
No description available.
Download README (.md)
Explore Similar Repositories
mobile_open
:
MLPerf Mobile benchmarks
eskiz-sms
:
python package for eskiz.uz/sms
unity-verify-code
:
An email parser to get 6 digit verification code
PiDiNet-NCNN
:
PiDiNet running in Android by ncnn
smace
:
Code for the paper "SMACE: A New Method for the Interpretability of Composite Decision Systems", ECML 2022
// repository documentation
Was this content helpful?
★ 0
(0 ratings)
Select Rating:
★
★
★
★
★
Submit Feedback
Recent Feedback
×
Download README
Do you want to download the
README.md
file for
CTNet
?
Download (.md)
# CTNet Official Pytorch implementation of the paper "[Contextual Transformation Network for Lightweight Remote Sensing Image Super-Resolution](https://ieeexplore.ieee.org/document/9632567)" accepted by IEEE TGRS. ## Requirements - Python 3.7 - Pytorch=1.5 - torchvision=0.6.0 - matplotlib - opencv-python - scipy - tqdm - scikit-image ## Installation Clone or download this code and install aforementioned requirements ``` cd codes ``` ## Dataset We used the UCMerced dataset for both training and testing. Please first download the dataset via [OneDrive](https://dlmueducn-my.sharepoint.com/:f:/g/personal/zyn2021_dlmu_edu_cn/EuwooOwtmepMrGgE_3x93mIB-3GKqBt0_4iOJ2mD8Xoj4A?e=EbiMOX) (key:912V). ## Download the results We share the super-resolved results generated by our CTNet. Then, researchers can compare their algorithms to our CTNet without performing inference. Results are available at [OneDrive](https://dlmueducn-my.sharepoint.com/:f:/g/personal/zyn2021_dlmu_edu_cn/EuwooOwtmepMrGgE_3x93mIB-3GKqBt0_4iOJ2mD8Xoj4A?e=EbiMOX) (key:912V). ## Train The train/val data pathes are set in [data/__init__.py](codes/data/__init__.py) ``` # x4 python demo_train_ctnet.py --model=CTNET --dataset=UCMerced --scale=4 --patch_size=192 --ext=img --save=CTNETx4_UCMerced # x3 python demo_train_ctnet.py --model=CTNET --dataset=UCMerced --scale=3 --patch_size=144 --ext=img --save=CTNETx3_UCMerced # x2 python demo_train_ctnet.py --model=CTNET --dataset=UCMerced --scale=2 --patch_size=96 --ext=img --save=CTNETx2_UCMerced ``` ## Test The test data path and the save path can be edited in [demo_deploy_ctnet.py](codes/demo_deploy_ctnet.py) ``` # x4 python demo_deploy_ctnet.py --model=CTNET --scale=4 # x3 python demo_deploy_ctnet.py --model=CTNET --scale=3 # x2 python demo_deploy_ctnet.py --model=CTNET --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](codes/metric_scripts/calculate_PSNR_SSIM.py) ``` cd metric_scripts python calculate_PSNR_SSIM.py ``` ## Citation If you find this work helpful, please consider citing the following paper: `````` @article{wang2022contextual, title={Contextual Transformation Network for Lightweight Remote Sensing Image Super-Resolution}, author={Wang, Shunzhou and Zhou, Tianfei and Lu, Yao and Di, Huijun}, journal={IEEE Transactions on Geoscience and Remote Sensing}, year={2022}, volume={60}, pages={1-13} } `````` ## Acknowledgements This code is built on [HSENet (Pytorch)](https://github.com/Shaosifan/HSENet). We thank the authors for sharing the codes.
