A Comparative Analysis of ECG-based Algorithms for Cardiac Arrhythmia Detection

Cardiac Arrhythmia detection has been the subject of considerable research in recent years, which is crucial for applications in medicine and sports. By analyzing the electrical signal of the heartbeat, which is created by distinctive, unique cardiac tissues situated in the body’s heart, these analyses assist in identifying many heart problems. Recently, a great deal of research has gone into creating class labels using automatic heartbeat categorization methods. The effectiveness of standard heart procedures has been improved using Deep Learning (DL) methods, which is particularly noteworthy. This research focuses on analyzing the ECG-based automatic detection of abnormal heartbeats using DL techniques. It examines the pre-processing of ECG signals to improve quality, segmentation of heartbeats to pinpoint the target region, extraction of features to simplify classifiers by reducing resource usage, and the use of different DL-based classification algorithms to generate labels for heartbeats. Many studies have shown that even small datasets can enhance classification performance. The study emphasizes the importance of improving the generalizability of DL models used for analyzing ECG signals, particularly by using larger datasets.