Hardware implementation of blind signal separation for efficient fetal PCG signal extraction
Manjusree Sreedharan Nair, Dhoulath Beegum Jalaludheen, Muhsina Noorjahan, Lubaib Panthar, Al Saheer Saheermanzil Subairkutty, Anjali AkkaraVeetil, Aswin Kunduparambil, Kiran Kishore, Allen Thomas Baiju · AIP conference proceedings · 2024
Blind source separation (BSS) is extracting a set of source signals from a group of mixed signals without knowledge about the original source signals or the mixing process.Nowadays, Blind Source Separation algorithms and different denoising methods are widely used in Biomedical Signal Processing applications.Fast ICA algorithm is one of the BSS algorithms that extract fetal heart sound from noisy single-channel abdominal phonocardiograms of pregnant women.This work implements the Savitzky-Golay Filtering method and Fast ICA BSS algorithm for signal enhancement on a portable device, Raspberry Pi, after the validation in Python and evaluates the performance of fetal Heart Sound extraction from mixed single channel abdominal PCG.The experiments on single channel abdominal PCG from pregnant women between 36 th to 40 th week of pregnancy are used to evaluate the methods.The results shows that this method extracts fetal heart sound from maternal abdominal PCG signal with a higher Signal-to-Noise ratio.