"Riemannian Geometric Computing: Applications in Imaging, Activity Modeling, and Mobile Health"

Pavan Turaga, Assistant Professor, Arizona State University

February 18th (Thursday), 3:00pm
Harold Frank Hall (HFH), Rm 4164

Many physical phenomenon — such as motion, shape, and light — when observed using traditional visual sensors, as well as novel emerging sensors such as depth cameras, orientation sensors, and smart personal devices, result in complex, high-dimensional spatiotemporal signatures that are difficult to analyze. The sources of difficulty are many: high data throughput, non-trivial transformations required for inference, physical variabilities such as temporal re-parametrization and sensor placement, and the non-Euclidean nature of feature spaces due to physical constraints on environments.

Over the past decade, many of these constraints have been found to be expressible in the language of Riemannian geometry. We will start with presenting several motivating examples from imaging, signal and sensor processing, where Riemannian computing plays an important role. Along the way we will present a brief overview of how one can extend classical multi-variate statistics, signal approximation theory, etc. to develop novel frameworks for characterizing visual (and non-visual) phenomenon. We will discuss Gauss-Markov processes as simple yet powerful models to describe the space of dynamical primitives in human movement analysis. We show how to use the Riemannian geometric properties of this primitive-space to devise effective inference algorithms, for applications in activity recognition and pattern discovery from long videos. We also consider geometrically meaningful representation of human movement as a product-space of non-Euclidean groups and function-spaces, and present algorithmic approaches for constructing statistical signal models on these spaces. We further outline the future potential of Riemannian geometric computing and the tools developed in this agenda for several impactful applications: including in physical activity based mobile-health systems.

About Pavan Turaga:

photo of Pavan Turaga Pavan Turaga is Assistant Professor jointly with the School of Arts, Media, Engineering, and Electrical Engineering at Arizona State University. He received the B.Tech. degree in electronics and communication engineering from the Indian Institute of Technology Guwahati, India, in 2004, and the M.S. and Ph.D. degrees in electrical engineering from the University of Maryland, College Park in 2008 and 2009 respectively. He then spent two years as a research associate at the Center for Automation Research, University of Maryland, College Park. His research interests are in signal processing and computer vision with current focus on differential geometric approaches. His works has applications in human activity analysis, dynamic scene analysis, and portable interventions for health and well-being. He is a Senior Member of the IEEE. He received the National Science Foundation's CAREER award in 2015, and the IEEE Phoenix Section Outstanding Faculty award in 2016. He was awarded the Distinguished Dissertation Fellowship in 2009 by the Univ. of Maryland, and was selected to participate in the Emerging Leaders in Multimedia Workshop by IBM, New York, in 2008.

Hosted by: B.S. Manjunath