Automated diagnostic tool for hypertension using convolutional neural network.

Area of Science:

• Biomedical Engineering
• Artificial Intelligence in Medicine
• Cardiology

Background:

• Hypertension (HPT) is a condition characterized by elevated blood pressure (BP) in arteries.
• Increased BP forces the heart to work harder, impacting cardiovascular health.
• Traditional 24-hour ambulatory blood pressure monitoring, while useful, presents practical challenges for patients.

Purpose of the Study:

• To develop an automated diagnostic tool for hypertension detection using electrocardiogram (ECG) signals.
• To overcome the inconvenience associated with conventional blood pressure monitoring methods.
• To leverage advanced signal processing and machine learning for efficient HPT diagnosis.

Main Methods:

• Electrocardiogram (ECG) signals were pre-processed.
• A convolutional neural network (CNN) model was employed to analyze ECG signatures.
• The model was validated using 10-fold and leave-one-out patient-based cross-validation techniques.

Main Results:

• The developed deep learning model achieved a high classification accuracy of 99.99% for detecting hypertension from ECG signals.
• Both 10-fold and leave-one-out validation methods confirmed the model's robust performance.
• This study represents an early application of deep learning for HPT detection via ECG.

Conclusions:

• The automated ECG-based tool demonstrates significant potential for effortless hypertension detection in hospital settings.
• This approach offers a promising, non-invasive method for routine HPT screening and diagnosis.
• The high accuracy suggests the clinical utility of deep learning in cardiovascular diagnostics.