Arrhythmia Detection and Classification on Cardiac Sensed Signals Using Deep Learning Techniques

Abstract

Arrhythmia is a general type of cardiac disease in persons between 30-40 years of age. Cardiac system in human body generates electrical pulses that can be captured and plotted through electrical system called ECG. Computer-aided diagnosis system (CADS) is a good approach to help the healthcare field for early, regular, and accurate diagnosis and treatment plan during critical care conditions. Deep learning-based CADS can helps in critical condition for more quick diagnosis and treatment in countries where doctor ratio is comparatively low. With the help of machine learning (ML) algorithm, intervariable relationships may be used for prediction. However, machine learning algorithms are also limited due to its datasets availability, established framework, and clinician unfamiliarity. This chapter aims to provide an idea of arrhythmia and CADS approach using cascaded deep learning model of CNN, LSTM, GRU, and RNN. The chapter focuses on techniques used in past years, comparative studies, and direction of research as future improvements in respective fields.