FOMO-MRI/FOMO60K

---
license: other
license_name: fomodataset
license_link: https://fomo25.github.io/dua
tags:
- brain
- mri
- ssl
- foundation_model
- 3d
- image
pretty_name: fomo-60k
size_categories:
- 10K<n<100K
task_categories:
- image-feature-extraction
- zero-shot-classification
---

# FOMO-60K: Brain MRI Dataset for Large-Scale Self-Supervised Learning with Clinical Data


_Dataset paper preprint:_ **A large-scale heterogeneous 3D magnetic resonance brain imaging dataset for self-supervised learning**.<br/>
[https://arxiv.org/pdf/2506.14432](https://arxiv.org/pdf/2506.14432).

This repo contains FOMO-60K, a large-scale dataset which includes MRIs from the clinic, containing 60K+ brain MRI scans.
This dataset is released in conjunction with the **FOMO25: Foundation Model Challenge for Brain MRI** hosted at **MICCAI 2025**.

See [fomo25.github.io](https://fomo25.github.io) for more information about the challenge.

Please be aware that usage of the dataset is subject to a [Data Usage Agreement](https://fomo25.github.io/dua) and a [Citation Policy](https://fomo25.github.io/citation-policy).

## Description

FOMO-60K is a large-scale dataset of brain MRI scans, including both clinical and research-grade scans. The dataset includes a wide range of sequences, including _T1, MPRAGE, T2, T2*, FLAIR, SWI, T1c, PD, DWI, ADC, and more_.

The dataset consists of
- 11,187 subjects
- 13,900 sessions
- 60,529 scans.

_For further details about the dataset, including the preprocessing, please consult the [dataset paper preprint](https://arxiv.org/pdf/2506.14432)_.

## Format

All data is provided as NIfTI-files. The dataset format is provided as one large standardized, preprocessed (including skull stripped, RAS reoriented, co-registered) dataset with the following format:

```
-- fomo-60k
|-- sub_01
|-- ses_01
|-- t1.nii.gz
```
Sessions with multiple scans of the same sequence are named `sequence_x.nii.gz`. Sessions where the sequence information was not available are named `scan_x.nii.gz`.

The dataset has been collected from the following public sources: OASIS, BraTS, MSD, IXI, MGH Wild, NKI, SOOP, NIMH, DLBS, IDEAS, ARC, MBSR, UCLA, QTAB, AOMIC ID1000.

许可证：其他
许可证名称：fomodataset
许可证链接：https://fomo25.github.io/dua
标签：
- 脑部
- 磁共振成像（MRI）
- 自监督学习（SSL）
- 基础模型（foundation model）
- 三维（3D）
- 图像
友好名称：FOMO-60K
样本量区间：10K<n<100K
任务类别：图像特征提取、零样本分类（zero-shot-classification）
---

# FOMO-60K：面向大规模自监督学习的脑部磁共振成像数据集（附临床数据）


**数据集预印本论文**：《面向自监督学习的大规模异构三维脑部磁共振成像数据集》<br/>
[https://arxiv.org/pdf/2506.14432](https://arxiv.org/pdf/2506.14432)

本仓库收录FOMO-60K——一款涵盖临床来源磁共振影像的大规模数据集，共计6万余张脑部MRI扫描影像。本数据集与在MICCAI 2025举办的**FOMO25：脑部磁共振成像基础模型挑战赛**同步发布。有关挑战赛的更多详情，请访问[fomo25.github.io](https://fomo25.github.io)。请注意，使用本数据集需遵守[数据使用协议](https://fomo25.github.io/dua)与[引用规范](https://fomo25.github.io/citation-policy)。

## 数据集描述

FOMO-60K是一款大规模脑部MRI扫描影像数据集，涵盖临床级与科研级两类扫描样本。该数据集包含多种成像序列，包括T1、MPRAGE、T2、T2*、FLAIR、SWI、T1c、PD、DWI、ADC等更多序列。

本数据集包含：
- 11187名受试者
- 13900次扫描会话
- 60529张扫描影像

如需了解数据集的更多细节（包括预处理流程），请查阅[数据集预印本论文](https://arxiv.org/pdf/2506.14432)。

## 数据格式

所有数据均以NIfTI文件格式提供。本数据集为统一标准化的预处理后数据集（包含颅骨剥离、RAS重定向、共配准对齐步骤），整体目录结构如下：

-- fomo-60k
|-- sub_01
|-- ses_01
|-- t1.nii.gz

若同一会话内存在同一序列的多张扫描影像，则将其命名为`sequence_x.nii.gz`；若无法获取序列信息，则将其命名为`scan_x.nii.gz`。

本数据集采集自以下公开数据源：OASIS、BraTS、MSD、IXI、MGH Wild、NKI、SOOP、NIMH、DLBS、IDEAS、ARC、MBSR、UCLA、QTAB、AOMIC ID1000。