A Novel Approach to Classify Electrocardiogram Signals Using Deep Neural Networks

Atrial fibrillation (AF) is an unusual heart rhythm condition which is caused by sporadic firing of electrical impulses from multiple places in the atria. AF is considered to be one of the leading cause of stroke in the world. AF is usually screened manually with the help of Electrocardiodiagram (ECG) reading. The manual process of reading ECG is a tedious and time-consuming task which is laden with human errors. Therefore, an automated process is quintessential. However, discerning anomaly in heart function using an efficient automated process has been a challenging task for quite some time. In this paper, the authors propose an intricate Neural Network architecture, CNN+LSTM, for the classification amongst four types of heart condition-Normal, Atrial Fibrillation, Noisy Sinus Rhythm and Alternative Rhythms using a dataset from PhysioNet/2017challenge. Volunteers in PhysioNet/2017 challenge dataset came from diverse backgrounds and had a wide window of variation in their physical attributes, making the dataset sufficiently reliable. In addition, the number of samples in this dataset exceeded any other before it on this topic, which further adds to the comprehensiveness of this dataset. The authors' method reached a summit accuracy of 91.19% using the proposed model.