ISIC 2017 Task 1: dermoscopy skin lesion segmentation, 2,750 images of melanoma, seborrheic keratosis and nevi with masks

Provider
International Skin Imaging Collaboration (ISIC); repackaged by MedOtter
License
CC0-1.0 (source)
Size
66 files
Format
Parquet (in sampled files)
Stored at
Hugging Face (https://huggingface.co/datasets/MedOtter/ISIC2017)
Access
Free and public. No account with the data provider is needed.
Category
Health & Biomedical
Source
https://doi.org/10.1109/ISBI.2018.8363547

About this dataset

Task 1 data from the ISBI 2017 challenge Skin Lesion Analysis Toward Melanoma Detection, hosted by the International Skin Imaging Collaboration (ISIC): dermoscopy photographs of skin lesions with expert binary lesion masks. There are 2,000 training, 150 validation and 600 test images (the final Test_v2 release), all with public ground truth, covering melanoma (521), seborrheic keratosis (386) and benign nevus (1,843). Resolutions range from about 540x722 to 4499x6748 pixels. Each row has the unmodified JPEG, the mask PNG (0 background, 255 lesion), approximate age and sex from the challenge metadata, and a flag for images that reappear in the ISIC 2018 training split; isic_2017_2018_split_ids.json lists the overlap. 63 Parquet files under data/, about 12 GB (measured 2026-11-21); read with Hugging Face datasets. Use for skin lesion segmentation and melanoma image analysis benchmarks. The ISIC challenge data page lists the 2017 data under CC-0. Cite Codella et al., ISBI 2018, doi 10.1109/ISBI.2018.8363547.

How to load this dataset in Python

from datasets import load_dataset
ds = load_dataset("MedOtter/ISIC2017")
print(ds)

Data structure

File types in sampled files: Parquet

Files

Example file: data/test-00000-of-00028.parquet

Analyze this dataset without downloading it

On Scigantic, ISIC 2017 Task 1: dermoscopy skin lesion segmentation, 2,750 images of melanoma, seborrheic keratosis and nevi with 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