Abstract:
Accurate extraction of the fetal QRS (FQRS) complex extracted from the maternal abdominal ECG (aECG) has become a very important clinical procedure to evaluate the well-being of the fetus. Fetal heart rate (FHR) obtained from the FQRS indicates fetal hypoxia, fetal distress, and other fetal-related problems that can be easily detected. In this paper, the aECG is preprocessed to remove the low-frequency baseline wandering and power line interference (PLI). In the following step, we proposed a technique to extract maternal QRS (MQRS) and FQRS from the Physionet abdominal ECG (aECG) database. The computed maternal and fetal QRS templates were optimized and fine-tuned for each of the Physionet noninvasive aECG databases. Our proposed synthesized QRS pulse template method was also compared with the Independent Component Analysis? (ICA) method for performance. The proposed system correctly evaluated and estimated the FECG signal for most records for the Physionet abdominal and direct fetal ECG database (adfecgdb) with the fetal QRS annotations. Using cross-correlation techniques, the MQRS and FQRS were estimated while the FHR was also correctly computed for most records.