# Real-time sleep disorder monitoring design using dynamic temporal graphs with facial and acoustic feature fusion

## TL;DR

This paper introduces a non-invasive system using facial and audio data to monitor sleep disorders in real-time, offering a more comfortable alternative to traditional methods.

## Contribution

A novel multimodal system using dynamic temporal graphs for real-time sleep disorder detection with high clinical accuracy.

## Key findings

- The system detected sleep apnea, restless leg syndrome, and cardiovascular irregularities with 94.6% clinical agreement.
- It achieved a 10.7-second average delay, suitable for real-time monitoring.
- The framework offers diagnostic accuracy comparable to traditional polysomnography.

## Abstract

Sleep disorders pose significant risks to patient safety, yet traditional polysomnography imposes substantial discomfort and laboratory constraints. We developed a non-invasive multimodal monitoring system for real-time sleep pathology detection.

We integrated facial expression analysis via deep convolutional neural networks with audio signal processing for breathing pattern detection. Heterogeneous data streams were unified into dynamic graph representations, with graph neural networks modeling spatiotemporal patterns of sleep pathologies.

The system accurately detected sleep apnea, restless leg syndrome, and cardiovascular irregularities with 10.7-s average delay and 94.6% clinical agreement, achieving diagnostic accuracy comparable to polysomnography.

This framework enables continuous non-invasive monitoring for point-of-care screening and home-based management, potentially expanding sleep medicine access for underserved populations.