Related Experiment Video
Updated: May 12, 2026
11:39
Non-invasive Optical Measurement of Cerebral Metabolism and Hemodynamics in Infants
Published on: March 14, 2013
20.2K
Artificial Intelligence non-invasive methods for neonatal jaundice detection: A review
Fati Oiza Salami1, Muhammad Muzammel1, Youssef Mourchid2
1Laboratoire Images, Signaux et Systémes Intelligents (LiSSi) EA 3956, Université Paris Est Créteil (UPEC), 122 Rue Paul Armangot, Vitry Sur Seine, Créteil, 94010, France.
Artificial Intelligence in Medicine
|February 23, 2025
Summary
Artificial intelligence (AI) offers a promising non-invasive method for diagnosing neonatal jaundice. AI models, including machine learning and deep learning, achieve over 90% accuracy, potentially reducing infant mortality.
Area of Science:
- Medical technology
- Artificial Intelligence in Healthcare
- Neonatal Medicine
Background:
- Neonatal jaundice is a significant cause of infant morbidity and mortality, particularly in low-resource settings.
- Traditional Total Serum Bilirubin (TSB) testing is invasive, posing risks of infection and delays.
- There is a critical need for accurate, non-invasive jaundice detection methods.
Purpose of the Study:
- To review and analyze non-invasive techniques for early neonatal jaundice detection.
- To evaluate the effectiveness and practicality of AI-driven diagnostic tools.
- To explore the potential impact of AI on reducing neonatal mortality.
Main Methods:
- Analysis of Artificial Intelligence (AI) techniques, including Machine Learning (ML) and Deep Learning (DL).
- Evaluation of AI models utilizing neonatal skin color and other features for jaundice detection.
- Review of mobile-based applications using smartphone cameras for bilirubin estimation.
Main Results:
- AI models, particularly neural networks, demonstrate over 90% accuracy in jaundice detection compared to traditional methods.
- Mobile applications show satisfactory field performance for estimating bilirubin levels in resource-limited areas.
- AI-based strategies are effective assistive tools for early jaundice diagnosis.
Conclusions:
- AI-powered non-invasive methods offer a revolution in neonatal care by enabling early jaundice diagnosis.
- These technologies present a practical alternative to invasive testing, especially in underserved regions.
- Further research into advanced imaging and wearable sensors is recommended for real-time monitoring.
