BUS-UCLM: 683 breast ultrasound images from 38 patients with lesion masks (benign, malignant, normal)

Provider
Noelia Vallez et al. (Hospital General de Ciudad Real); mirrored by MedOtter
License
CC-BY-4.0 (source)
Size
3 files
Format
Parquet (in sampled files)
Stored at
Hugging Face (https://huggingface.co/datasets/MedOtter/BUS-UCLM)
Access
Free and public. No account with the data provider is needed.
DOI
10.17632/7fvgj4jsp7.3
Category
Health & Biomedical
Source
https://doi.org/10.17632/7fvgj4jsp7.3

About this dataset

BUS-UCLM is a breast ultrasound lesion segmentation dataset collected at Hospital General de Ciudad Real, Spain, on a Siemens ACUSON S2000 with the 18L6 HD probe. It holds 683 images from 38 patients: 174 with benign lesions, 90 malignant and 419 normal, at roughly 856x606 pixels. The original masks are RGB color-coded by class (green benign, red malignant); this mirror stores one binary lesion mask per image and keeps the class in a class_label column, alongside patient_id and native resolution, in a Parquet file under data/. There is no official split; the authors' helper repository suggests a 33/5 patient partition. About 330 MB (measured 2026-11-17). Use for breast lesion segmentation and benign versus malignant classification in ultrasound. Mirrored on Hugging Face by MedOtter from the authors' release, with a card documenting provenance and any normalization. The authors' Mendeley Data record (doi 10.17632/7fvgj4jsp7) is CC BY 4.0; cite Vallez et al., Scientific Data 2025.

How to load this dataset in Python

from datasets import load_dataset
ds = load_dataset("MedOtter/BUS-UCLM")
print(ds)

Data structure

File types in sampled files: Parquet

Files

Example file: data/train-00000-of-00001.parquet

Analyze this dataset without downloading it

On Scigantic, BUS-UCLM: 683 breast ultrasound images from 38 patients with lesion masks (benign, malignant, normal) 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