We provide our PyTorch implementation of structure-preserving image translation using a mutual information loss to enforce depth consistency. This code is adapted from the implementation of Contrastive Unpaired Translation (CUT) with the addition of mutual information loss based on the PyTorch Lightning implementation of mutual information score.
Our sample scripts 'train.sh' and 'predict.sh' to demonstrate how to run the training/inference phases. Please see the original CUT repository for more in-depth instructions.
If you use our code for your research, please cite our paper.
@inproceedings{wang2024structpres,
title={Structure-preserving Image Translation for Depth Estimation in Colonoscopy},
author={Shuxian Wang and Akshay Paruchuri and Zhaoxi Zhang and Sarah McGill and Roni Sengupta},
booktitle={Medical Image Computing and Computer Assisted Intervention},
year={2024}
}
Our code is adapted from Contrastive Unpaired Translation (CUT).