Cuff-less Blood Pressure Measurement Based on Deep Convolutional Neural Network.

Area of Science:

• Biomedical Engineering
• Cardiovascular Health
• Artificial Intelligence in Medicine

Background:

• Cuff-less blood pressure (BP) monitoring is crucial for managing cardiovascular events.
• Existing methods often rely on manual feature extraction, limiting their ability to capture complex physiological signal-BP relationships.

Purpose of the Study:

• To develop and evaluate a deep learning-based method for cuff-less BP estimation.
• To leverage automatic feature learning from raw physiological signals for improved BP monitoring.

Main Methods:

• Utilized a 16-layer VGGNet architecture to process electrocardiogram (ECG) and pressure pulse wave (PPW) signals.
• Employed automatic feature learning capabilities of deep networks, eliminating the need for manual feature extraction.
• Evaluated the method on 89 middle-aged and elderly subjects, comparing results against oscillometric BP measurements.

Main Results:

• Achieved a high correlation coefficient of 0.91 for systolic BP and 0.89 for diastolic BP.
• Reported estimation errors of -2.06 ± 6.89 mmHg for systolic BP and -4.66 ± 4.91 mmHg for diastolic BP.
• Demonstrated commendable accuracy in cuff-less BP estimation.

Conclusions:

• The proposed VGGNet-based method offers a promising approach for accurate cuff-less BP estimation.
• Automatic feature learning from ECG and PPW signals provides a novel insight into BP monitoring.
• This technology has the potential to significantly aid in cardiovascular event management.