A multi-modal approach using a non-parametric model to extract fetal ECG

Saman Noorzadeh (1, 2) , Bertrand Rivet (2) , Pierre-Yves Guméry (1)

1 TIMC-IMAG - Techniques de l'Ingénierie Médicale et de la Complexité - Informatique, Mathématiques et Applications, Grenoble - UMR 5525
2 GIPSA-VIBS - GIPSA - Vision and Brain Signal Processing

This study presents a non-parametric method to extract and separate maternal and fetal Electrocardiogram (ECG) from an abdominal channel. The proposed method relies on the use of two additional reference signals related to the maternal and to the fetal ECGs. The abdominal recording fetal and maternal ECGs are modeled from their respective references using a Gaussian process. Both of these signals are modeled considering a thoracic ECG reference, and an abdominal PCG reference, respectively for maternal ECG and fetal ECG. This general method provides a good behavior to extract fetal ECG as stated by the numerical experiments and shows that it can use only 1-bit reference signals, which can be easier to record and more efficient to process.

Mots clés: multi-modal; Gaus-sian process; Index Terms— fetal ECG; non-parametric model; PCG; multi-channel approach

Domaines: Traitement du signal et de l'image [eess.SP]

Saman Noorzadeh, Bertrand Rivet, Pierre-Yves Guméry. A multi-modal approach using a non-parametric model to extract fetal ECG. ICASSP 2015 - 40th IEEE International Conference on Acoustics, Speech and Signal Processing, Apr 2015, Brisbane, Australia. ⟨10.1109/ICASSP.2015.7178086⟩. ⟨hal-01208418⟩