Video-based detection of cessation of breathing in pre-term infants using machine learning

2026-07-06

AI Summary

This study addresses the challenge of apnea detection in preterm infants, who are particularly vulnerable due to immature respiratory control. Conventional contact-based monitoring is often compromised by motion artifacts, sensor displacement, and skin fragility. To overcome these limitations, the authors propose a non-contact video-based monitoring approach that dynamically tracks the thoracic region to extract respiratory time-series signals. They develop a multimodal deep learning model based on ResNet, integrating video-derived respiration with physiological signals including impedance pneumography, electrocardiogram, and photoplethysmography-derived respiration. The study demonstrates, for the first time, that video signals alone can detect apnea events with a balanced accuracy of 76.9%. When fused with physiological modalities, performance significantly improves to 90.6%, outperforming any single-modality approach and offering a more robust and information-rich solution for neonatal respiratory monitoring.

Abstract

Pre-term infants are susceptible to potentially harmful apnoea-related cessations of breathing due to immature respiratory control. However, reliable respiratory monitoring in the neonatal intensive care unit (NICU) remains challenging because motion artefacts, sensor displacement, and skin fragility can compromise contact-based measurements. Non-contact video monitoring offers a complementary approach that does not depend on adhesive sensors while providing additional respiratory information. We investigated whether camera-based signals can detect apnoea-related cessation of breathing (COBE) and provide complementary information to routinely acquired physiological signals. Using video and clinical recordings from 30 pre-term infants, respiratory motion was extracted from dynamically tracked torso regions to generate camera-derived time-series signals. Camera-only models were trained using residual network (ResNet) architectures, while hybrid models combined video-derived signals with impedance pneumography (IP), ECG-derived respiration (EDR), and the PPG-derived respiratory envelope. Camera-only models achieved a balanced accuracy of 76.9%, demonstrating the feasibility of non-contact COBE detection. Combining video-derived features with IP improved balanced accuracy to 90.6%, outperforming either modality alone and indicating that video provides respiratory information beyond standard physiological signals. These findings show that video-derived signals contain clinically relevant respiratory features and enhance COBE detection when combined with conventional physiological signals. This supports non-contact video as a complementary modality for automated COBE detection and highlights its potential to improve the robustness of neonatal respiratory monitoring.

Problem

• apnoea
• cessation of breathing
• pre-term infants
• respiratory monitoring
• non-contact video

Innovation

• non-contact video monitoring
• apnoea detection
• respiratory motion extraction
• hybrid multimodal fusion
• ResNet-based modeling