Covariance Tracking using Model Update Based on Means on Riemannian Manifolds.

Fatih Porikli (1), Oncel Tuzel (1,2) and Peter Meer (2,3)

(1) Mitsubishi Electric Research Laboratories
Cambridge, MA 02139
(2) Department of Computer Science
(3) Department of Electrical and Computer Engineering
Rutgers University, Piscataway, NJ 08854, USA

We propose a simple and elegant algorithm to track non-rigid objects using a covariance based object description and an update mechanism based on means on Riemannian manifolds. We represent an object window as the covariance matrix of features, therefore we manage to capture the spatial and statistical properties as well as their correlation within the same representation. The covariance matrix enables efficient fusion of different types of features and modalities, and its dimensionality is small. We incorporated a model update algorithm using the elements of Riemmanian geometry. The update mechanism effectively adapts to the undergoing object deformations and appearance changes. The covariance tracking method does not make any assumption on the measurement noise and the motion of the tracked objects, and provides the global optimal solution. We show that it is capable of accurately detecting the non-rigid, moving objects in non-stationary camera sequences while achieving a promising detection rate of 97.4 percent.

2006 Computer Vision and Pattern Recognition Conference , New York City, NY, June 2006, vol. I, 728-735.
Return to Research: Content-based Retrieval        Return to List of Publications
Download the paper