Feature Learning Using Convolutional Neural Network for Cardiac Arrest Detection

Abstract

Arrhythmias including ventricular fibrillation and ventricular tachycardia, which are known as shockable rhythms, are the mainly cause of sudden cardiac arrests (SCA). In this paper, we propose a feature learning scheme applied for detection of SCA on electrocardiogram signal with the modified variational mode decomposition technique. The subsequent SAA consists of a convolutional neural network as a feature extractor (CNNE) and a support vector machine classifier. The features extracted by selected CNNE are then validated using 5-folds CV procedure on the evaluation data, and enable the accuracy of 99.02 %, sensitivity of 95.21 %, and specificity of 99.31 %.