This study deals with fetal ECG extraction by multi-modal non-parametric modeling. In a recently proposed non-parametric approach, the fetal and maternal ECG are modeled from their respective references using Gaussian processes. The reference signals for maternal ECG and fetal ECG are the thoracic ECG and abdominal PCG, respectively. In this paper, multiple kernels are used instead of a single kernel for the covariance function of the Gaussian process. The proposed methodology improves the results of fetal ECG extraction as confirmed by experimental results.

Additional Metadata
Keywords Fetal ECG, Gaussian processes, Multi-channel, Multi-kernel, Multimodal, Non-parametric model, PCG
Persistent URL dx.doi.org/10.1109/GlobalSIP.2017.8309099
Conference 5th IEEE Global Conference on Signal and Information Processing, GlobalSIP 2017
Citation
Surisetti, B. (Bharathi), & Dansereau, R. (2018). Multi-modal fetal ECG extraction using multi-kernel Gaussian processes. In 2017 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2017 - Proceedings (pp. 943–947). doi:10.1109/GlobalSIP.2017.8309099