Quasi-Periodic Non-negative Matrix Factorization for Phonocardiographic signals denoising

Nafissa Dia (1) , Julie Fontecave-Jallon (1) , Pierre-Yves Guméry (2) , Bertrand Rivet (3)

1 TIMC-IMAG - Techniques de l'Ingénierie Médicale et de la Complexité - Informatique, Mathématiques et Applications, Grenoble - UMR 5525
2 TIMC-IMAG-PRETA - Physiologie cardio-Respiratoire Expérimentale Théorique et Appliquée
3 GIPSA-VIBS - GIPSA - Vision and Brain Signal Processing

Mechanical cardiac activity may be monitored with phonocardiographic (PCG) signals giving access to cardiac sounds. However, many noises interfere with cardiac information in raw signals and denoising such signals is necessary before interpretation. Non-negative Matrix Factorization (NMF) is of interest for time-frequency representations to separate noise and signal components. In this paper, to exploit the quasi-periodicity of the PCG, a quasi-periodic NMF (QP-NMF), based on multiplicative updates derived from a Majoration-Minimization algorithm, is proposed to decompose the PCG spectrograms. Numerical simulations show the good behavior of the proposed method to separate quasi-periodic components from the others. Finally, applied on real noisy PCG signals, QP-NMF shows its interest compared to an unsupervised NMF to denoise PCG signals.

Domains: Signal and Image processing; Imaging

Nafissa Dia, Julie Fontecave-Jallon, Pierre-Yves Guméry, Bertrand Rivet. Quasi-Periodic Non-negative Matrix Factorization for Phonocardiographic signals denoising. SAM 2018 - 10th IEEE Workshop on Sensor Array and Multichannel Signal Processing, Jul 2018, Sheffield, United Kingdom. pp.390-394. ⟨hal-01847123⟩