Real-Time Cuffless Continuous Blood Pressure Estimation Using 1D Squeeze U-Net Model: A Progress toward mHealth

Area of Science:

• Biomedical Engineering
• Artificial Intelligence in Healthcare
• Cardiovascular Monitoring

Background:

• Continuous blood pressure (BP) monitoring is crucial for healthcare advancement.
• Current cuff-based methods are inconvenient for real-time, continuous measurement.
• Mobile health (mHealth) offers potential for remote patient monitoring.

Purpose of the Study:

• To propose a real-time, cuffless method for measuring blood pressure (BP) using a smartphone.
• To develop an energy-efficient, 1D Squeeze U-net model for BP estimation from photoplethysmogram (PPG) signals.
• To create an Android application for convenient home BP measurement and hypertension assessment.

Main Methods:

• Utilized a 1D Squeeze U-net deep learning model to process raw photoplethysmogram (PPG) signals.
• Assessed the model's accuracy and reliability on 100 individuals across MIMIC-I and MIMIC-III datasets.
• Developed a smartphone application for real-time, cuffless BP estimation.

Main Results:

• The model achieved high accuracy in estimating systolic BP (MAE 4.42 mmHg), diastolic BP (MAE 2.25 mmHg), and mean arterial pressure (MAE 2.56 mmHg).
• Results met the British Hypertension Society's grade A performance requirements and AAMI error range.
• Demonstrated that short PPG signal segments are sufficient for accurate real-time BP measurement.

Conclusions:

• A novel, real-time cuffless BP estimation method using mHealth and deep learning is presented.
• The developed system enables accurate home BP measurement and supports hypertension assessment.
• This approach advances continuous cardiovascular monitoring through accessible mobile technology.