Hau-Tieng Wu, Turning nonstationary biomedical signals into useful clinical information by denoising manifolds

Submitted by Ingrid Richter on

The Department of Applied Mathematics weekly seminar is given by scholars and researchers working in applied mathematics, broadly interpreted. 

 


Title: Turning nonstationary biomedical signals into useful clinical information by denoising manifolds

Abstract: In the clinical arena we are moving beyond snapshot health data. Physicians are now provided and confronted by multimodal physiological data collected over long stretches of time, and some are of low quality or contaminated by undesired information. The nonstationarity and heterogeneity nature of these datasets can impose a serious challenge for health care providers and medical researchers, particularly when they need clinically useful and actionable information at the bedside. I will discuss recent progress in signal processing dealing with some of these challenges. The main tool is circling around the mission called manifold denoising, and our solution depends on random matrix theory and spectral geometry. Clinical examples and current progress will be structured toward providing practical solutions.

Share