Modern people commonly face sleep disturbances that affect health and quality of life. As the global sleep market continues to expand, it reflects the increasing public attention to sleep problems. Among these, sleep apnea is a condition that requires special attention. However, traditional sleep testing methods remain limited to hospital settings, requiring the attachment of multiple physiological sensors and consuming substantial medical manpower and time, significantly reducing patients' willingness to seek medical care.

To overcome these limitations, our hospital's Sleep Medicine Center and Artificial Intelligence Center have collaborated successfully to develop the " Home-based Sleep Detection System"! The most distinctive feature of this innovative system is: it requires only a single-lead ECG sensor to automatically generate complete sleep reports through AI technology, including the critical AHI index for diagnosing sleep apnea. Patients can complete testing in the familiar home environment and upload data to the cloud in real-time, allowing AI models to quickly analyze and assist physicians in remote tracking and evaluation. This not only enhances the comfort and acceptability of the testing process but also creates a new era of precision medicine!

Patients only need to collect compact devices from the sleep center to conduct continuous sleep monitoring for up to 7 days in the comfort of their homes. The testing process is simple and convenient: through an intuitive app, patients can easily upload data to our hospital's AI cloud platform after waking up to generate preliminary inference results, which are then reviewed by professional sleep technicians, ultimately providing physicians with accurate diagnoses and customized treatment recommendations.

The system's device is not only lightweight and easy to wear but also greatly reduces sleep disturbances, allowing patients to fall asleep with peace of mind and increasing testing willingness. The system also has superior battery life for continuous monitoring, providing physicians with the most complete diagnostic basis. The innovative hardware design significantly reduces testing costs, making comprehensive sleep monitoring affordable for more patients. Additionally, eliminating trips to the hospital allows patients to complete testing in familiar home environments, truly realizing convenient, reassuring precision medicine!

This system not only automatically calculates AHI (apnea-hypopnea index), precisely determining the severity of sleep apnea, but also detects complete sleep stages, helping physicians comprehensively understand patients' sleep conditions. Through fully automated analysis, it significantly reduces technicians' manual marking time, effectively lowering medical costs. The current system performance is as follows: mild or above sleep apnea (AHI>5) 87.91%, moderate or above (AHI>15) 89.58%, severe (AHI>30) 95.83%, apnea event determination AUC 0.96, sleep-wake event determination AUC 0.91. From April 27, 2023, to April 15, 2025, there have been 3,303 accumulated uses, helping to screen out individuals with sleep disorders. 

Journal Article

• EfficientNet-based machine learning architecture for sleep apnea identification in clinical single-lead ECG signal data sets. 
   https://doi.org/10.1186/s12938-024-01252-w

​Awards

2023

• NHQA National Healthcare Quality Award
• SNQ Symbol of national Quality

Patents

• Taiwan patent certification (Certificate No. I815546)

Contact window

Project Management Group
Phone: +886-4-22052121 ext 12534