Fetal Arrhythmia Detection Using Fetal ECG Signal

Sudden infant death syndrome (SIDS) has remained a challenge to overcome for the medical practitioner. Among other causes, the fetal arrhythmia is accountable for a significant portion of such cases. Any heart rate of a baby above 160 bpm or below 120 bpm refers to fetal arrhythmia. In comparison with various diagnostic methodology, ECG is a low-cost non-invasive method which measures the electrical activity of the heart. Thus, to detect fetal arrhythmia, we developed an ECG signal feature extracting algorithm and extracted eight significant features of the fetal ECG signal. Based on these features, Kernel Support Vector Machine (SVM) classifier with Gaussian Kernel was utilised to detect fetal arrhythmia. For evaluating the learning model, we used the leave one out (LOO) cross-validation. The final result displayed accuracy of 83.33% with 91.67% specificity and 75% sensitivity. Thus, this research shows a way of developing a unique non-invasive and low-cost fetal arrhythmia diagnosis method.