---
license: apache-2.0
tags:
  - medical-imaging
  - brain-segmentation
  - 3d-unet
  - unet
  - mri
  - uhf-mri
  - contrast-agnostic
  - resolution-agnostic
pipeline_tag: image-segmentation
---

# GOUHFI 2.0

This repository hosts the model weights for **GOUHFI 2.0**, a 3D U-Net-based deep learning framework for brain segmentation, cortical parcellation and volumetry measurements using Magnetic Resonance Images (MRI) of any contrast, resolution or field strength.

## Source Code

For the full source code, preprocessing pipeline, training scripts, and inference instructions, please visit the official repository available on GitHub:

https://github.com/mafortin/GOUHFI

## Archival Release

The official archival release of the trained model weights is available on Zenodo:

https://zenodo.org/records/17920473

## Paper

If you use this work, please cite:

- GOUHFI original publication in Imaging Neuroscience:
```bibtex
@article{fortin2025gouhfi,
  title={GOUHFI: A novel contrast-and resolution-agnostic segmentation tool for ultra-high-field MRI},
  author={Fortin, Marc-Antoine and Kristoffersen, Anne Louise and Larsen, Michael Staff and Lamalle, Laurent and Stirnberg, R{\"u}diger and Goa, P{\aa}l Erik},
  journal={Imaging Neuroscience},
  volume={3},
  pages={IMAG--a},
  year={2025}
}
```

- Pre-print of GOUHFI 2.0:
```bibtex
@article{fortin2026gouhfi,
  title={GOUHFI 2.0: A Next-Generation Toolbox for Brain Segmentation and Cortex Parcellation at Ultra-High Field MRI},
  author={Fortin, Marc-Antoine and Kristoffersen, Anne Louise and Goa, Paal Erik},
  journal={arXiv preprint arXiv:2601.09006},
  year={2026}
}
```

## Intended Use

This model is intended for research use only.

It is not intended for clinical diagnosis, treatment planning, or medical decision-making without appropriate validation and regulatory approval.

## License

Apache License 2.0