Computational Mathematics and Statistics Seminar by Whitney Huang: Modeling High Frequency Oscillating Biomedical Signals Using Synchrosqueezing Transform and Locally Stationary Gaussian Processes

Speaker: Whitney Huang, Assistant Professor, School of Mathematics and Statistics, Clemson University

Title: Modeling High Frequency Oscillating Biomedical Signals Using Synchrosqueezing Transform and Locally Stationary Gaussian Processes

Abstract: I will introduce a new class of models, Synchrosqueezing Transform and local stationary Gaussian Processes, to relate several oscillating biomedical signals. Especially to properly extract important oscillatory features, a nonlinear-type time-frequency analysis tool, the synchrosqueezing transform, is employed; these time-frequency features are then used to estimate the oscillating biomedical signal of interest by a locally stationary Gaussian process regression. I will mainly focus on an application for estimating airflow signals from thoracic and abdominal movement signals, which may have some merits in the context of sleep telemedicine. This is joint work with Yu-Min Chung, Hau-Tieng Wu (Duke), and Yu-Bo Wang (Clemson). 

Contact Yindong Chen for joining details.

Computational Mathematics and Statistics Seminar