Particle Filtering on Riemannian Manifolds. Application to Visual Tracking

Speaker

Hichem Snoussi, Ph.D, HdR

 

Full Professor,

Charles Delaunay Institute, UMR STMR 6279, CNRS

University of Technology of Troyes, France

 

Talk 2:

Title: Particle Filtering on Riemannian Manifolds. Application to Visual Tracking

VenueRoom 406, No.5 SEIEE Buiding

Time: 10:00AM, June 14,Tuesday

Abstract: 

Recently, a general scheme of particle filtering on Riemannian manifolds has been proposed in  [Snoussi 2006]. In addition to the nonlinear dynamics, the system state is constrained to lie on a Riemannian manifold, which dimension is much lower than the whole embedding space dimension. The Riemannian manifold formulation of the state space model  avoids the curse of dimensionality from which suffer most of the particle filter methods. Furthermore, this formulation is the only natural tool when the embedding Euclidean space cannot be defined (the state space is defined in an abstract geometric way) or when the constraints are not easily handled (space of positive definite matrices).  In order to  illustrate the effectiveness of the proposed differential-geometric framework, we consider the problem of visual tracking. The specificity of our modeling is the extension of the hidden state (velocity and position) by jointly estimating the state covariance.  As the state covariance is a positive definite matrix,  the Euclidean space is not suitable when tracking this covariance. Instead, one should  exploit the differential geometric properties of the space of positive definite matrices, by constraining the estimated matrix to move along the geodesics of this Riemannian manifold.   The proposed sequential Bayesian updating consists thus  in drawing state samples while moving on the manifold geodesics.

 

Speaker Biography:

Hichem Snoussi was born in Bizerta, Tunisia, in 1976. He received the diploma degree in electrical engineering from the Ecole Superieure d'Electricite (Supelec), Gif-sur-Yvette, France, in 2000. He also received the DEA degree and the Ph.D. in signal processing from the University of Paris-Sud, Orsay, France, in 2000 and 2003 respectively. He has obtained the HdR from the University of Technology of Compiègne in 2009. 

Between 2003 and 2004, he was postdoctoral researcher at IRCCyN, Institut de Recherches en Communications et Cybernétiques de Nantes. He has spent short periods as visiting scientist at the Brain Science Institute, RIKEN, Japan and Olin Neuropsychiatry Research Center at the Institute of Living in USA. Between 2005 and 2010, he has been associate professor at the University of Technology of Troyes. Since September 2010, he has been appointed a Full Professor position at the same university. He is in charge of the regional research program S3 (System Security and Safety) of the CPER 2007-2013 and the CapSec plateform (wireless embedded sensors for security). He is the principal investigator of an ANR-Blanc project (mv-EMD), a CRCA project (new partnership and new technologies) and a GDR-ISIS young researcher project. He is partner of many ANR projects, GIS, strategic UTT programs. He obtained the national doctoral and research supervising award PEDR 2008-2012.  

[ 2011-06-14 ]