IDRiD: Indian Diabetic Retinopathy Image Dataset, 81 fundus photographs with microaneurysm, hemorrhage, exudate and optic disc masks

Provider
Prasanna Porwal et al. (IDRiD organizers); repackaged by MedOtter
License
CC-BY-4.0 (source)
Size
8 files
Format
Parquet (in sampled files)
Stored at
Hugging Face (https://huggingface.co/datasets/MedOtter/IDRiD)
Access
Free and public. No account with the data provider is needed.
DOI
10.21227/H25W98
Category
Health & Biomedical
Source
https://doi.org/10.3390/data3030025

About this dataset

The lesion segmentation part of IDRiD (Indian Diabetic Retinopathy Image Dataset), released for the ISBI 2018 Diabetic Retinopathy: Segmentation and Grading challenge with images from India. It holds 81 color fundus photographs of the retina (54 training, 27 test) with pixel-level masks for the lesions of diabetic retinopathy: microaneurysms (MA), hemorrhages (HE), hard exudates (EX) and soft exudates or cotton-wool spots (SE), plus the optic disc (OD). Not every image contains every lesion type, so some mask columns are empty; a lesion_types column lists which are present. Six Parquet files under data/ with images and masks as PNG bytes, about 0.5 GB (measured 2026-11-21); read with Hugging Face datasets. The grading and localization parts of IDRiD are not included. Use for retinal lesion segmentation and diabetic eye disease screening models. Original data on IEEE DataPort (DOI 10.21227/H25W98), released under CC BY 4.0. Cite Porwal et al., Data 3(3):25 (2018), doi 10.3390/data3030025.

How to load this dataset in Python

from datasets import load_dataset
ds = load_dataset("MedOtter/IDRiD")
print(ds)

Data structure

File types in sampled files: Parquet

Files

Example file: data/test-00000-of-00002.parquet

Analyze this dataset without downloading it

On Scigantic, IDRiD: Indian Diabetic Retinopathy Image Dataset, 81 fundus photographs with microaneurysm, hemorrhage, exudate and optic disc masks 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