Skip to content
Back to all 100 research datasets
Medical ImagingCredentialed / DUA required

CheXpert

Curated by Stanford Machine Learning Group / Stanford AIMI

Resource Overview

Large chest radiograph benchmark with uncertainty-aware labels derived from radiology reports and expert reference annotations.

Cohort & Sample Size

224,316 radiographs; 65,240 patients

Data Modalities & Features

Chest X-rayRadiologyImage ClassificationReports

Data Governance & Compliance Notice

Credentialed access requires verified researcher identity and signed Data Use Agreement (DUA) through the official curator repository. Trace does not host or mirror patient-level data.

Access Repository

Curator / InstitutionStanford Machine Learning Group / Stanford AIMI
License / TermsStanford CheXpert Research Use Agreement
Access ClassificationCredentialed / DUA required

More in Medical Imaging

NIH ChestX-ray14

NIH Clinical Center

Open access
MIMIC-CXR

MIT Laboratory for Computational Physiology / BIDMC

Credentialed / DUA required
MIMIC-CXR-JPG

MIT Laboratory for Computational Physiology

Credentialed / DUA required