Related Experiment Video
Updated: Aug 8, 2026
06:27
Transcutaneous Microcirculatory Imaging in Preterm Neonates
Published on: December 31, 2015
NewbornTimeLine: Automated Video-Based Timelines for Neonatal Resuscitation in a Hospital Pilot Study
IEEE Journal of Biomedical and Health Informatics
|August 6, 2026
Summary
NewbornTimeLine uses AI and video analysis to automatically create accurate timelines of neonatal resuscitation events. This system improves documentation accuracy and supports quality improvement research in clinical settings.
Area of Science:
- Medical Informatics
- Artificial Intelligence in Healthcare
- Neonatal Care Technology
Background:
- Manual documentation of neonatal resuscitation is time-consuming and lacks precision.
- Accurate event timelines are crucial for quality improvement and research in neonatal care.
Purpose of the Study:
- To develop and evaluate NewbornTimeLine, an AI-based system for automated generation of neonatal resuscitation timelines.
- To assess the feasibility of AI-driven documentation in a clinical setting.
Main Methods:
- Utilized thermal and visible light video recordings for automated analysis.
- Developed a two-stream fusion architecture for time-of-birth detection (93.80% accuracy).
- Employed ROI-centered MoViNet for Neonatal Resuscitation Algorithm (NRA) activity recognition (F1-scores of 0.97 and 0.77).
Main Results:
- Achieved 100% birth identification and high accuracy in time-of-birth detection.
- Demonstrated strong performance in recognizing key resuscitation activities like ventilation and stimulation.
- Successfully piloted the system in a single hospital, proving feasibility of automated timeline generation.
Conclusions:
- Automated timeline generation for neonatal resuscitation using AI is feasible in a clinical setting.
- NewbornTimeLine offers a promising tool for retrospective analysis, clinical debriefing, and quality improvement.
- The system enhances the accuracy and efficiency of documenting critical neonatal events.
