Validation of a deep learning model for classification of pediatric pneumonia in Hong Kong
• A deep learning model was developed to screen pediatric chest x-rays for pneumonia. • The model automatically classifies chest x-rays as positive or negative for pneumonia. • Model was validated by comparing its output to classification from trained evaluators. • Model has high accuracy, sensitivity and specificity compared to human interpretation. • Cropping the images improved the model's classification performance.
Authors
- Tsz K Khan
- Jessica Weaver (ORCID: https://orcid.org/0000-0002-9861-3328)
- Elaine Lee (ORCID: https://orcid.org/0000-0002-0627-5297)
- Craig S. Roberts (ORCID: https://orcid.org/0000-0002-5852-9770)
- Kate Ching Ching Chan (ORCID: https://orcid.org/0000-0001-7272-8285)
- Meghan White (ORCID: https://orcid.org/0000-0002-1215-116X)
- Boshu Ru (ORCID: https://orcid.org/0000-0001-9620-1306)
- Kim S. J. Lao (ORCID: https://orcid.org/0000-0002-8488-0708)
- Dong Wang (ORCID: https://orcid.org/0000-0002-0545-6205)
- Yiyun Chen
- Andy Cheuk Nam Hwang
Institutions
- Merck & Co., Inc., Rahway, NJ, USA (United States) (US)
- Chinese University of Hong Kong (HK)
- Asia Centre (TH)
- University of Hong Kong (HK)
Publication Details
- Journal
- Vaccine
- Published
- 2024-09-21
- DOI
- https://doi.org/10.1016/j.vaccine.2024.126370
- Citations
- 2
- Primary Topic
- Pneumonia and Respiratory Infections
- Type
- article
- Field-Weighted Citation Impact
- 0.77