tuvovan / rfdnet-imagesuperresolution

Keras Implementation of the paper Residual Feature Distillation Network for Lightweight Image Super-Resolution

RFDNet Super Resolution

Residual Feature Distillation Network for Lightweight Image Super-Resolution

Content

• RFDNet
• Getting Started
• Running
• References
• Citations

Getting Started

• Clone the repository

Prerequisites

• Tensorflow 2.2.0+
• Python 3.6+
• Keras 2.3.0
• PIL
• numpy

pip install -r requirements.txt

Running

Training

• Train RFDNet

  python main.py
  

• Test RFDNet

  python test.py
  

Usage

Testing

usage: test.py [-h] [--test_path TEST_PATH] [--gpu GPU]
               [--weight_test_path WEIGHT_TEST_PATH] [--filter FILTER]
               [--feat FEAT] [--scale SCALE]

optional arguments:
                    -h, --help            show this help message and exit
                    --test_path TEST_PATH
                    --gpu GPU
                    --weight_test_path WEIGHT_TEST_PATH
                    --filter FILTER
                    --feat FEAT
                    --scale SCALE

Result

Input - Low Res	Bilinear	Output High Res

License

This project is licensed under the MIT License - see the LICENSE file for details

References

[1] Training and Testing dataset - link

Citation

@misc{liu2020residual,
      title={Residual Feature Distillation Network for Lightweight Image Super-Resolution}, 
      author={Jie Liu and Jie Tang and Gangshan Wu},
      year={2020},
      eprint={2009.11551},
      archivePrefix={arXiv},
      primaryClass={eess.IV}
}

Acknowledgments

• Any ideas on updating or misunderstanding, please send me an email: [email protected]
• If you find this repo helpful, kindly give me a star.

rfdnet-imagesuperresolution's Issues

Weights for 'scale == 2' Model

Hello @tuvovan

Thank you for this TF implementation and work !

I would appreciate it if you could kindly please share the weights (h5 files) for a x2 scale please?

TypeError: 'Image' object is not iterable

when running test.py
File "test.py", line 20, in run
out, out_bilinear = upscale_image(model, lr)
TypeError: 'Image' object is not iterable

NameError: name 'test_img_paths' is not defined

Traceback (most recent call last):
File "main.py", line 91, in
callbacks = [ESPCNCallback(), early_stopping_callback, model_checkpoint_callback]
File "/content/drive/MyDrive/Experiment/codes/36_RFDN/RFDNet-ImageSuperResolution/utils.py", line 96, in init
self.test_img = get_lowres_image(load_img(test_img_paths[0]), upscale_factor)
NameError: name 'test_img_paths' is not defined

How can I feed my dataset (have LR and HR respectively) into the program?

Hello tuvovan.
Thank you for your effort.
Indeed, I have a dataset that contain HR and LR pair.
May I know how can I input the the images into your program to train?
Thank you very much.

there is not CCA in RFDB

I wonder if you remove cca 0.0

PIL not found

No matching distribution found for PIL==8.0.0