Covariance Tracking via Geometric Particle Filtering

Region covariance descriptor recently proposed has has been approved robust and elegant to describe a region of interest, which has been applied to visual tracking.The covariance matrix enables efficient fusion of different types of features, where the spatial and statistical properties as well as their correlation are characterized. The similarity of two covariance descriptor is measured on Riemannian manifolds. Within a probabilistic framework, we integrate covariance descriptor into Monte Carlo tracking technique for visual tracking. Most existing particle filtering based tracking algorithms treat deformation parameters of the target as a vector. We have proposed a visual tracking algorithm via geometric particle filtering, which implements the particle filter with the constraint that the system state lies in a low dimensional manifold: affine lie group. The sequential Bayesian updating consists in drawing state samples while moving on the manifold geodesics; Theoretic analysis and experimental evaluations against the tracking algorithm based on geometric particle filtering demonstrate the promise and effectiveness of this algorithm.