LIMUC: 11,276 colonoscopy images of ulcerative colitis labeled with Mayo endoscopic score, 564 patients (Marmara University)

Provider
Gorkem Polat, Haluk Tarik Kani, Ilkay Ergenc et al. (Marmara University); repackaged by Chehab Lab
License
CC-BY-4.0 (source)
Size
5 files
Format
Parquet (in sampled files)
Stored at
Hugging Face (https://huggingface.co/datasets/chehablab/LIMUC)
Access
Free and public. No account with the data provider is needed.
DOI
10.5281/zenodo.5827695
Category
Health & Biomedical
Source
https://doi.org/10.1093/ibd/izac226

About this dataset

LIMUC (Labeled Images for Ulcerative Colitis) is a colonoscopy image dataset for grading the severity of ulcerative colitis, a chronic inflammatory bowel disease. It holds 11,276 endoscopic images from 564 patients examined at Marmara University School of Medicine, Istanbul, each graded by experienced gastroenterologists with the Mayo endoscopic subscore: Mayo 0 normal or inactive (6,105), Mayo 1 mild (3,052), Mayo 2 moderate (1,254), and Mayo 3 severe (865). This copy is built from the authors' patient-based release, so every row carries a patient ID for patient-level evaluation; images were converted losslessly from BMP to PNG at native resolution and an integer label added. Three Parquet files under data/, about 1.3 GB (measured 2026-11-21); read with Hugging Face datasets. Use for ulcerative colitis severity classification and inflammatory bowel disease endoscopy models. Original data on Zenodo (DOI 10.5281/zenodo.5827695), released under CC BY 4.0. Cite Polat et al., Inflammatory Bowel Diseases 29(9):1431-1439 (2023), doi 10.1093/ibd/izac226.

How to load this dataset in Python

from datasets import load_dataset
ds = load_dataset("chehablab/LIMUC")
print(ds)

Data structure

File types in sampled files: Parquet

Files

Example file: data/train-00000-of-00003.parquet

Analyze this dataset without downloading it

On Scigantic, LIMUC: 11,276 colonoscopy images of ulcerative colitis labeled with Mayo endoscopic score, 564 patients (Marmara University) opens in a hosted JupyterLab notebook with the data already mounted, so you can read files where they are stored instead of copying them first. Each notebook includes an AI assistant that can see the mounted files and write and run analysis code. Open the dataset page to launch a notebook, or browse the full catalog.

This page is also available as Markdown and as JSON.

Related Health & Biomedical datasets