A Review on Heartbeat Classification for Arrhythmia Detection Using ECG signal Processing

The electrocardiogram (ECG) provides essential characteristics of the human heart's multiple cardiac conditions. The classification of arrhythmias provides a major part in the diagnosis of cardiac disease. Any deviation from the normal sequence of electrical impulses is considered an arrhythmia. Traditional methods of signal processing, machine learning and its sub-branches, such as deep learning, are popular techniques for ECG signal analysis and classification and, above all, for the development of early detection and treatment applications for cardiac conditions and arrhythmias. This article presents a detailed literature survey on ECG signal analysis. This paper aims to analyze the most recent studies on data utilized, features, and machine learning approaches that can address the time computational challenge and be implemented in wearable technology. The study methodology began with a search for relevant papers, followed by a study of the data provided. The second stage was to explore the evaluated ECG characteristics and the machine learning method used to identify arrhythmia. According to the analysis, a significant number of studies selected the MIT-BIH database, even though it needs a substantial ratio of pre-processing effort. We address a detailed existing research work review on the data of real-time signal collection, pre-recorded diagnostic ECG data, analysis and denoising of ECG signals, identification of ECG spectrographic states based upon function technologies, and classification of ECG signals, as well as comparative discussions between the studies analyzed.