A Novel Approach to Visualize Arrhythmia Classification Using 1D CNN

Cardiac-related disorders have been one of the major concerns in recent decades. The electrocardiogram, an extensively utilized medical instrument, records the electrical activity of the heart as a wave. Cardiac arrhythmia is a condition of having an irregular heartbeat. Manually identifying irregularities in an ECG wave is a complicated and challenging task. The current work focuses on computationally identifying the ECG wave fluctuations to determine the abnormality in the heartbeat. We propose to use a 1-Dimensional Convolutional Neural Network (CNN) that analyses a given ECG signal data to identify irregularities in the functioning of the heart and represent the associated risks using graphics interchange format (GIF) files of a 3-dimensional heart. We obtained an accuracy score of 96.72% in classifying given ECG data into five different arrhythmia classes. Automated detection and visual representation of cardiac conditions can help medical associates easily interpret ECG signals and determine arrhythmia early.