Non-invasive Single Channel integration model for fetal ECG extraction and sustainable fetal healthcare using wavelet framework

A retrospective aspect of prenatal complexities during pregnancy and advancements in technology shows the need for unscathed fetal ECG extraction from a single mother abdominal ECG (abdECG). The proposed work introduces a Non-invasive Single Channel Integration Technique (NSCIT) depicting a cumulative trapezoidal mathematical model with an LMS adaptive algorithm for mother and fetal ECG extraction with improved Signal-to-Noise Ratio (SNR). Besides separation, fetal ECG (fECG) features are extracted, simulated, analyzed, and compared with standards to generate fetus cardiac growth during later Gestation Period (GP) of 21st to 40th week of pregnancy. The variants of the wavelet transform, such as Dual-Tree Complex Wavelet Transform (DTCWT) for pre-processing and Maximal Overlap Discrete Wavelet Transform (MODWT) for post-processing, are exploited using a multi-resolution analysis. The NSCIT algorithm with LMS adaptive technique has shown 100% accuracy for detecting mother ECG and specific fetal ECG extraction channels. The improved accuracy using abdominal lead 4 is 96.36%, and overall abdominal mixed lead accuracy is 93.32% compared with recent existing literature. The maximum error in comparing Power Spectral Density (PSD) of actual and extracted fECG and mECG is significantly less. The calculated correlation coefficient between actual and extracted fetal QRS width, fetal R-peak intervals (R-R), and fetal heart rate (fHR) for Db1 are 0.70, 0.99, and 0.67, respectively. The research outcomes show that fECG SNR increases with GP, and it is maximum for the GP of 40th week. This fECG morphological analysis before childbirth will efficaciously contribute to sustainable fetal healthcare.