Cardio Sense: CNN - BiGRU Powered Deep Learning for Heart Sound Analysis

Abstract

Since Cardiac and vascular conditions rank among major cause of death around the world, sophisticated diagnostic methods are required for both early identification and prevention [13] Libby et al. In order to predict heart functionality with high accuracy, we present a novel deep learning model in this study that uses heart acoustic inputs. Due to the hybrid conventional neural network (CNN) and architecture with a gate with a gate with a re-unit (BIGRU) architecture, the recommended model learns the time and spatial features of heart sound records. In contrast to conventional machine learning methods, approach improves a small diagnosis of heart abnormalities with improved diagnostic accuracy.