Fetal electrocardiogram extraction and enhancement using triggered adaptive filtering

Analysis of the electrocardiogram (ECG) is a simple and noninvasive method of effectively assessing a patient's general health and diagnosing cardiovascular diseases. The ECG can also be used to clinically evaluate the health of prenatal infants. Variations in the FECG waveform morphology can provide important diagnostic information that is not available in current fetal monitoring techniques, such as ultrasound and phonocardiography. Such evaluations offer the opportunity to combat prenatal health conditions sooner, and offer prenatal awareness of potential physical abnormalities and complications.; The objective of this thesis is to investigate a novel multichannel and multimodal FECG extraction (stage 1) and enhancement (stage 2) technique using linear adaptive filtering with triggering, based on noninvasive abdominal measurements. In stage 1, we show that by controlling the periods of adaptation, a more accurate FECG signal can be extracted. The concept of triggering is proposed with the inclusion of an extra modality, such as the fetal heart sound or Doppler ultrasound data, to be combined with the electrical modality of the ECG, to provide temporal information about the FECG. Using synthetic ECG data, our results indicate a 5.49% PPRD improvement using triggering with linear transitions compared to no triggering. Slight improvements were obtained with triggering compared to no triggering with simulated FECG data and real data.; In stage 2, we investigate the extension of triggering to FECG enhancement using multiple abdominal channels. Our results indicate an increase in the SNRest of 2.61 using triggering with linear transitions over no triggering with simulated FECG data, and an increase of 0.81 using triggering with on/off transitions over no triggering with real data.