# Figshare brain tumor MRI dataset (Cheng et al.): meningioma, glioma and pituitary tumor

## Facts

- Provider: Jun Cheng et al., Southern Medical University, Guangzhou (converted by MedOtter)
- License: CC-BY-4.0 (https://figshare.com/articles/dataset/brain_tumor_dataset/1512427)
- Size: 6 files
- Format: Parquet (in sampled files)
- Stored at: Hugging Face, https://huggingface.co/datasets/MedOtter/Figshare_Brain_Tumor
- Access: public, readable without an account
- DOI: https://doi.org/10.6084/m9.figshare.1512427.v8
- Category: Health & Biomedical
- Source: https://doi.org/10.1371/journal.pone.0140381
- Page: https://scigantic.com/archive/fef2925a-0b57-5c1d-bfbe-196f1f0437b3
- Keywords: Health & Biomedical, MRI, image segmentation, benchmark, Parquet, Hugging Face

## About

This is the widely used brain tumor dataset of Cheng et al., published on Figshare with their PLoS ONE 2015 paper on tumor retrieval and classification. It contains 3,064 T1-weighted contrast-enhanced MRI slices from 233 patients with three tumor types: meningioma (708 slices), glioma (1,426) and pituitary tumor (930). Every slice comes with a binary tumor mask and the original hand-delineated tumor border. The MedOtter conversion stores each slice as a 512 x 512 16-bit PNG with its mask, border polyline, patient ID, label and the five patient-disjoint cross-validation folds of the original release; fold 1 is used as test (542 slices) and folds 2 to 5 as train (2,522). As measured 2026-11-14 the data are four Parquet files totalling about 0.7 GB; load with datasets (load_dataset('MedOtter/Figshare_Brain_Tumor')) or pyarrow. A standard benchmark for brain tumor classification and segmentation with CNNs and transformers. The authors release the data on Figshare under CC BY 4.0; cite Cheng et al. 2015.

## Load the data

```python
from datasets import load_dataset
ds = load_dataset("MedOtter/Figshare_Brain_Tumor")
print(ds)
```

## Data structure

File types in sampled files: Parquet

Example file: `data/test-00000-of-00001.parquet`

## Open it

- In a hosted notebook with the data mounted: https://scigantic.com/archive/fef2925a-0b57-5c1d-bfbe-196f1f0437b3
- Full record as JSON (schema card included): https://api.scigantic.com/api/archives/fef2925a-0b57-5c1d-bfbe-196f1f0437b3
- Over MCP: `get_schema_card` with id `fef2925a-0b57-5c1d-bfbe-196f1f0437b3` at https://api.scigantic.com/mcp

## Related topics

- [MRI datasets](https://scigantic.com/topics/mri)
- [Parquet datasets](https://scigantic.com/topics/parquet)
- [Benchmark datasets](https://scigantic.com/topics/benchmark)
- [Image segmentation datasets](https://scigantic.com/topics/image-segmentation)
- [Health & Biomedical datasets](https://scigantic.com/topics/health-and-biomedical)
