Related Experiment Video
Updated: Sep 16, 2025
09:24
Quantified Assessment of Infant's Gross Motor Abilities Using a Multisensor Wearable
Published on: May 17, 2024
1.6K
Evaluating Non-Invasive Computer Vision-Based Quantification of Neonatal Movement as a Marker of Development in
Janet Pigueiras-Del-Real1, Lionel C Gontard1,2, Isabel Benavente-Fernández3,4,5
1Department of Condensed Matter Physics, University of Cádiz, 11510 Cádiz, Spain.
Healthcare (Basel, Switzerland)
|July 12, 2025
Summary
Computer vision objectively quantifies preterm infant movement using contactless pose tracking. This method offers a reliable, complementary indicator of neonatal development alongside traditional anthropometric measures.
Area of Science:
- Neonatal development
- Biomedical engineering
- Computer vision applications
Background:
- Traditional neonatal assessments use anthropometric measures (weight, size).
- Objective movement quantification is emerging as a complementary indicator for preterm infants.
Purpose of the Study:
- Evaluate computer vision for non-invasive neonatal movement quantification.
- Identify reliable algorithms and optimal recording duration for pose tracking.
Main Methods:
- Utilized contactless pose tracking with computer vision.
- Analyzed ~800,000 postural data points from ten preterm infants.
- Compared whole-body vs. regional tracking and various algorithms (optical flow, distance-based).
Main Results:
- 30-second video segments suffice for consistent motion quantification.
- Distance-based algorithms (Chebyshev, Minkowski) showed high stability (CV 5.46%, 6.40%).
- Lower-body tracking with Minkowski and Mahalanobis metrics closely matched full-body results (0.89%, 1% difference).
Conclusions:
- Computer vision enables objective, contactless quantification of neonatal movement.
- This approach can complement traditional measures for monitoring preterm infant progression.
- Potential applications include continuous monitoring and early clinical decision-making.
