UIEPTA

Dataset Preparation

Download UIEB Dataset and EUVP dataset

Folder Structure

• uw_data/: This is the main folder that contains all the underwater image datasets.

• train/: This folder is where all the training data is stored.

  • a/: This folder holds the input images, which are the degraded underwater images used to train your model.

  • b/: This folder contains the reference images, also known as ground truth images. These images are used to evaluate how well the model performs during training.

• test/: This folder contains all the testing data.

  • a/: This folder has the input images that will be used to test the trained model.

  • b/: This folder holds the reference images that correspond to the test input images. These images are used to check the quality of the model's outputs.

Training Instructions

To train the model, execute the following command in your terminal:

python train.py

Citation

If you found this work UIEPTA helpful, please cite it using the following reference:

@inproceedings{khan2023underwater,
  title={Underwater image enhancement with phase transfer and attention},
  author={Khan, MD Raqib and Kulkarni, Ashutosh and Phutke, Shruti S and Murala, Subrahmanyam},
  booktitle={2023 International Joint Conference on Neural Networks (IJCNN)},
  pages={1--8},
  year={2023},
  organization={IEEE}
}

For any inquiries or further information, please contact me at srk689800@gmail.com.