Detection of Fetal Cardiac Anomaly from Composite Abdominal Electrocardiogram

Detecting the presence of fetal cardiac anomaly from Fetal Electrocardiogram (FECG) is intricate and challenging, but it is vital to know the health status of the fetus. This paper develops a Fetal Cardiac Anomaly Detection (FCAD) system that focuses on FECG signal extraction and identification of pathologic fetus using clinically essential features hidden in the amplitudes and waveform durations of the FECG signals. The proposed FCAD system consists of six main stages: (i) Abdominal Electrocardiogram (AECG) signal Acquisition (ii) Preprocessing (iii) FECG extraction (iv) Post-processing (v) Feature extraction and (vi) Pathologic fetus characterization using Support Vector Machine (SVM) classifier. Three approaches based on Fast Independent Component Analysis (Fact ICA) algorithm, Waveform Correspondence Algorithm (WCA) and Smoothed Pseudo Wigner-Ville Distribution (SPWVD), a time-frequency based technique were employed for FECG extraction and their performances were assessed based on the computed performance measures such as Signal-to-Noise Ratio (SNR), Root Mean Square Error (RMSE) and Correlation Coefficient (rho).The performance of the FCAD system was evaluated on AECG signals taken from publicly available databases and real signals. Classification results evince that the features derived from FECG signals obtained using SPWVD gave promising best accuracy 97% than others. The experimental results suggest that the developed FCAD system has enormous potential and promise in the early detection of pathologic fetuses.

Automated summary

This study developed a FCAD system focusing on extracting FECG signals and identifying pathological fetuses using SVM classifier. The performances of three extraction methods were evaluated based on SNR, RMSE, and rho, with the SPWVD method showing the highest accuracy.