23 January 2017 Single-channel mixed signal blind source separation algorithm based on multiple ICA processing
Author Affiliations +
Proceedings Volume 10322, Seventh International Conference on Electronics and Information Engineering; 1032203 (2017) https://doi.org/10.1117/12.2265268
Event: Seventh International Conference on Electronics and Information Engineering, 2016, Nanjing, China
Abstract
Take separating the fetal heart sound signal from the mixed signal that get from the electronic stethoscope as the research background, the paper puts forward a single-channel mixed signal blind source separation algorithm based on multiple ICA processing. Firstly, according to the empirical mode decomposition (EMD), the single-channel mixed signal get multiple orthogonal signal components which are processed by ICA. The multiple independent signal components are called independent sub component of the mixed signal. Then by combining with the multiple independent sub component into single-channel mixed signal, the single-channel signal is expanded to multipath signals, which turns the under-determined blind source separation problem into a well-posed blind source separation problem. Further, the estimate signal of source signal is get by doing the ICA processing. Finally, if the separation effect is not very ideal, combined with the last time’s separation effect to the single-channel mixed signal, and keep doing the ICA processing for more times until the desired estimated signal of source signal is get. The simulation results show that the algorithm has good separation effect for the single-channel mixed physiological signals.
© (2017) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Xiefeng Cheng, Ji Li, "Single-channel mixed signal blind source separation algorithm based on multiple ICA processing", Proc. SPIE 10322, Seventh International Conference on Electronics and Information Engineering, 1032203 (23 January 2017); doi: 10.1117/12.2265268; https://doi.org/10.1117/12.2265268
PROCEEDINGS
9 PAGES


SHARE
Back to Top