Skip to content

MOrphing histoPAthology DIffusion models for explainability of AI pipelines in medical image analysis.

License

Notifications You must be signed in to change notification settings

KatherLab/mopadi

Repository files navigation

MoPaDi - Morphing Histopathology Diffusion

MoPaDi combines Diffusion Autoencoders with multiple instance learning (MIL) for explainability of deep learning classifiers in histopathology. This repository contains the supplementary material for the following preprint:

Žigutytė, L., Lenz, T., Han, T., Hewitt, K. J., Reitsam, N. G., Foersch, S., ... & Kather, J. N. (2024). Counterfactual Diffusion Models for Mechanistic Explainability of Artificial Intelligence Models in Pathology. bioRxiv, 2024.

image info

For segmentation of 6 cell types to quantify changes in original and counterfactual images, DeepCMorph pretrained models were used.

Getting started

Create a virtual environment, e.g. with conda or mamba, clone the repository, and install required packages:

mamba create -n mopadi python=3.8 -c conda-forge
pip install -r requirements.txt

Pretrained Models

TBA

Datasets

Acknowledgements

This project was built upon a DiffAE (MIT license) repository. We thank the developers for making their code open source.

Reference

If you find our work useful for your research or if you use parts of the code please consider citing our preprint:

Žigutytė, L., Lenz, T., Han, T., Hewitt, K. J., Reitsam, N. G., Foersch, S., ... & Kather, J. N. (2024). Counterfactual Diffusion Models for Mechanistic Explainability of Artificial Intelligence Models in Pathology. bioRxiv, 2024.
@misc{zigutyte2024mopadi,
      title={ounterfactual Diffusion Models for Mechanistic Explainability of Artificial Intelligence Models in Pathology}, 
      author={Laura Žigutytė and Tim Lenz and Tianyu Han and Katherine Jane Hewitt and Nic Gabriel Reitsam and Sebastian Foersch and Zunamys I Carrero and Michaela Unger and Alexander T Pearson and Daniel Truhn and Jakob Nikolas Kather},
      year={2024},
      eprint={2024.10.29.620913},
      archivePrefix={bioRxiv},
      url={https://www.biorxiv.org/content/10.1101/2024.10.29.620913v1}, 
}

About

MOrphing histoPAthology DIffusion models for explainability of AI pipelines in medical image analysis.

Resources

License

Stars

Watchers

Forks

Releases

No releases published