Building a software to detect arrhythmia and extract fetal ECG from material ECG automatically

Abstract

Electrocardiogram (ECG) is an electrical signal containing information about the condition and functioning of the heart. Nowadays, many types of arrhythmias can be efficiently diagnosed based on this signals. In this study, we developed a software based on Matlab GUI to analyze ECG data recorded from the ECG 9620 Nihon Kohden device. The software read data, calculated Heart Rate, detected and analyzed PR, RR, ST intervals and width of QRS. Algorithms were developed to identify arrhythmia. In this study, Fast Independent Component Analysis (Fast ICA) method is also utilized to extract fetal ECG (FECG) from maternal ECG and calculate heart rate of fetal. The software is also very friendly to users and allowed medical doctors and staff to work and analyze those data conveniently. We tested the software with over 300 ECG data and FECG data obtained from simulators and patients monitored by experts and medical doctors. The software could recognize 12 types of common arrhythmia types with good precision and help users to analyze fetal ECG with high accuracy without invasive measurement method. These results indicated that our software is useful to support medical staff to detect arrhythmias and monitor FECG.