Overview

Schema Version

https://atlas.rsna.org/schemas/2025-11/dataset.json

Name

ChestX-ray8 (also referred to as ChestX-ray14)

Link

https://nihcc.app.box.com/v/ChestXray-NIHCC/folder/36938765345

Indexing

Keywords: ChestX-ray8, ChestX-ray14, chest radiographs, dataset, natural language processing labels, weakly supervised

Content: CH

Author(s)

National Institutes of Health Clinical Center

Summers RM

Organization(s)

National Institutes of Health Clinical Center

Comments

ChestX-ray8 is a large chest radiograph dataset created and made publicly available by the NIH Clinical Center; originally labeled for 8 diseases and later expanded to 14 (often referred to as ChestX-ray14). Labels in the original dataset were extracted from radiology reports using NLP.

Date

Published: 2017

References

[1] Wang X, Peng Y, Lu L, Lu Z, Bagheri M, Summers RM. "ChestXray8: hospital-scale chest-ray database and benchmarks on weakly supervised classification and localization of common thorax diseases v5". 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). 2017.

[2] National Institutes of Health. "NIH clinical center provides one of the largest publicly available chest x-ray datasets to scientific community". . . Available from: https://www.nih.gov/news-events/news-releases/nih-clinical-center-provides-one-largest-publicly-available-chest-x-ray-datasets-scientific-community

[3] National Institutes of Health Clinical Center. Summers R, editor.. "ChestX-ray8; 2017". . . Available from: https://nihcc.app.box.com/v/ChestXray-NIHCC/folder/36938765345

Dataset

Motivation

Large-scale chest radiograph dataset to enable training and benchmarking of ML models for thoracic disease classification/localization.

Missing information

The article does not provide exact counts, demographics, file formats, partitions, licensing terms, or detailed acquisition/preprocessing protocols.