PyTorch implementation of the paper "EAGLE: An Edge-Aware Gradient Localization Enhanced Loss for CT Image Reconstruction". This repository includes the code for our novel Eagle-Loss function, designed to improve the sharpness of reconstructed CT image.

The Eagle_Loss code is developed using Python 3.11 and PyTorch 2.0.0. To ensure compatibility, please install the necessary packages using the following commands to create and activate a conda environment:

conda env create -f environment.yml
conda activate eagle_loss

FOV extension data can be downloaded here.

This repository is organized as follows:

• dataset.py: This script is responsible for handling the dataset.

• eagle_loss.py: Contains the implementation of the Eagle-Loss function. For patch_size, we suggest set to 3.

• model.py: Defines the architecture of the U-Net that is used for FOV extension.

• train.py: This script is used to train the model.

• Training script.
• Pre-trained model weights.
• Usage examples.

@article{sun2024eagle,
  title={EAGLE: An Edge-Aware Gradient Localization Enhanced Loss for CT Image Reconstruction},
  author={Sun, Yipeng and Huang, Yixing and Schneider, Linda-Sophie and Thies, Mareike and Gu, Mingxuan and Mei, Siyuan and Bayer, Siming and Maier, Andreas},
  journal={arXiv preprint arXiv:2403.10695},
  year={2024}
}