README.md

---
datasets:
- AnonRes/OpenMind
license: cc-by-4.0
pipeline_tag: image-segmentation
tags:
- medical
---

# OpenMind Benchmark 3D SSL Models

This repository hosts pre-trained checkpoints from the **OpenMind** benchmark, including the **Primus** architecture.

> **Models from the papers**: 
> - [Primus: Enforcing Attention Usage for 3D Medical Image Segmentation](https://huggingface.co/papers/2503.01835)
> - [An OpenMind for 3D medical vision self-supervised learning](https://arxiv.org/abs/2412.17041)
> **Pre-training codebase**: [MIC-DKFZ/nnssl](https://github.com/MIC-DKFZ/nnssl)  
> **Downstream (segmentation) framework**: [TaWald/nnUNet](https://github.com/TaWald/nnUNet) / [MIC-DKFZ/nnUNet](https://github.com/MIC-DKFZ/nnUNet)
> **Dataset**: [AnonRes/OpenMind](https://huggingface.co/datasets/AnonRes/OpenMind)

---

![OpenMind](https://huggingface.co/datasets/AnonRes/OpenMind/resolve/main/assets/OpenMindDataset.png)

## Overview

This repository provides self-supervised pre-trained weights for 3D medical image analysis. These models were pre-trained on the [OpenMind Dataset](https://huggingface.co/datasets/AnonRes/OpenMind), a large-scale collection of brain MRI data.

**Primus** and **PrimusV2** are Transformer-centric segmentation architectures designed to maximize the effectiveness of attention mechanisms in 3D medical imaging. By moving away from heavy convolutional reliance, Primus achieves state-of-the-art results on several benchmarks.

**These models are not recommended to be used as-is for feature extraction.** Instead, we recommend using the downstream fine-tuning frameworks for **segmentation** available in the [adaptation repository](https://github.com/TaWald/nnUNet).

---

## Model Variants

We release SSL checkpoints for two primary backbone architectures:

- **ResEnc-L**: A CNN-based encoder [[a](https://arxiv.org/abs/2410.23132), [b](https://arxiv.org/abs/2404.09556)]  
- **Primus-M**: A transformer-based encoder introduced in the [Primus paper](https://huggingface.co/papers/2503.01835)

Each encoder has been pre-trained using various SSL techniques:

| Method        | Description |
|---------------|-------------|
| [Volume Contrastive (VoCo)](https://arxiv.org/abs/2402.17300) | Contrastive pretraining method for 3D volumes |
| [VolumeFusion (VF)](https://arxiv.org/abs/2306.16925)         | Spatial volume fusion-based segmentation SSL method |
| [Models Genesis (MG)](https://www.sciencedirect.com/science/article/pii/S1361841520302048) | Reconstruction and denoising based pretraining method |
| [Masked Autoencoders (MAE)](https://openaccess.thecvf.com/content/CVPR2022/html/He_Masked_Autoencoders_Are_Scalable_Vision_Learners_CVPR_2022_paper) | Default reconstruction based pretraining method |
| [Spark 3D (S3D)](https://arxiv.org/abs/2410.23132) | Sparse reconstruction based pretraining method (CNN only) |
| [SimMIM](https://openaccess.thecvf.com/content/CVPR2022/html/Xie_SimMIM_A_Simple_Framework_for_Masked_Image_Modeling_CVPR_2022_paper.html) | Simple masked reconstruction based pretraining method (TR only) |
| [SwinUNETR SSL](https://arxiv.org/abs/2111.14791) | Rotation, Contrastive and Reconstruction based pre-training method. |
| [SimCLR](https://arxiv.org/abs/2002.05709) | Transfer of 2D Contrastive learning baseline method to 3D |

## Usage

To use these models for segmentation, please refer to the [nnU-Net documentation for Primus](https://github.com/MIC-DKFZ/nnUNet/blob/master/documentation/primus.md).

```bash
pip install nnunetv2
```

## Citation

If you use these models, please cite the following papers:

```bibtex
@article{wald2025primus,
  title={Primus: Enforcing Attention Usage for 3D Medical Image Segmentation},
  author={Wald, Tassilo and Roy, Saikat and Isensee, Fabian and Ulrich, Constantin and Ziegler, Sebastian and Trofimova, Dasha and Stock, Raphael and Baumgartner, Michael and K{\"o}hler, Gregor and Maier-Hein, Klaus},
  journal={arXiv preprint arXiv:2503.01835},
  year={2025}
}

@article{wald2024openmind,
  title={An OpenMind for 3D medical vision self-supervised learning},
  author={Wald, Tassilo and Ulrich, Constantin and Suprijadi, Jonathan and Ziegler, Sebastian and Nohel, Michal and Peretzke, Robin and K{\"{o}}hler, Gregor and Maier-Hein, Klaus H},
  journal={arXiv preprint arXiv:2412.17041},
  year={2024}
}
```