Lung Ultrasound Imaging Dataset for Accurate Detection and Localization of LUS Vertical Artifacts

Published: 25 September 2025| Version 2 | DOI: 10.17632/h923m366sf.2
Contributors:
,
,
,
,
,
,
,
,
,

Description

This dataset contains 401 high-resolution lung ultrasound (LUS) images annotated with polygonal bounding boxes identifying LUS Vertical artifacts. The images were collected from 255 patients with pulmonary diseases at Mulago and Kiruddu National Referral Hospitals in Uganda. LUS Vertical artifacts are important indicators for conditions like pulmonary edema, interstitial lung disease, pneumonia, and COVID-19. This dataset provides a resource for training and validating deep learning models for B-line detection and localization, helping to improve AI-assisted respiratory diagnostics.

Files

Institutions

Makerere University

Categories

Medicine, Computer Vision, Ultrasound, Deep Learning

Funding

International Development Research Centre

Swedish International Development Cooperation Agency

Makerere University

Licence