Cardiac Arrhythmia Detection from Single-lead ECG using CNN and LSTM assisted by Oversampling

Abstract: Cardiac Arrhythmia is one of most commonly occurring cardiovascular disease in which atrial fibrillation is most commonly occurring arrhythmia which can be detected using single lead electrocardiogram. In this research, we develop a Deep Learning (DL) model with a combination of Convolutional Neural Net and Long Short-term Memory assisted by Oversampling technique which classifies the 2017 PhysioNet/CinC Challenge dataset into four classes, i.e. normal sinus rhythm, atrial fibrillation, others and noisy classes with an accuracy better than present techniques. We can integrate this Algorithm to CPS-heart to find abnormalities in the human heart.  

Algorithm used for Classification.

Download Paper