A New Approach to Classify Cardiac Arrythmias Using 2D Convolutional Neural Networks.

Area of Science:

• Cardiology
• Artificial Intelligence
• Biomedical Signal Processing

Background:

• Cardiovascular diseases represent a leading global cause of mortality.
• Early detection of cardiovascular diseases, including arrhythmias, is crucial for effective treatment and reduced mortality.
• Electrocardiogram (ECG) signals are vital for diagnosing heart conditions.

Purpose of the Study:

• To propose and validate a new methodology for detecting 17 types of arrhythmias using 2D Convolutional Neural Networks (CNNs).
• To assess the efficacy of using 15x15 pixel grayscale images of ECG heartbeat segments for arrhythmia detection.

Main Methods:

• Development of a novel detection system employing 2D Convolutional Neural Networks.
• Transformation of ECG signal heartbeat segments into 15x15 pixel grayscale images for input into the CNN.
• Utilizing the MIT-BIH arrhythmia database for model training and validation.

Main Results:

• The proposed methodology achieved a precision of 92.31% in detecting arrhythmias.
• The performance is comparable to the current state-of-the-art in arrhythmia detection.
• The system demonstrated effectiveness in identifying a wide range of 17 distinct arrhythmias.

Conclusions:

• The study presents a novel, automated method for arrhythmia detection from ECG signals.
• The 2D CNN approach using image representations of heartbeats shows significant promise for clinical application.
• This methodology offers a valuable tool for early and accurate diagnosis of cardiovascular conditions.