2004 Research Highlights

This page highlights some of our recently published research results:

IROS 04: Inference In The Space Of Topological Maps: An MCMC-based Approach
Ananth Ranganathan and Frank Dellaert. Using MCMC sampling to perform inference about the topology of a robot's environment.

IROS 04: Map-Based Priors for Localization
Sang Min Oh, Sarah Tariq, Bruce Walker, and Frank Dellaert. Using previously available maps about surface use as a prior in localizing humans or robots.

CVPR 04: A Rao-Blackwellized Particle Filter for EigenTracking
Zia Khan, Tucker Balch, Frank Dellaert. An efficient method for using subspace representations in a particle filter by applying Rao-Blackwellization to integrate out the subspace coefficients.

CVPR 04: Atlanta World: An Expectation-Maximization Framework for Simultaneous Low-level Edge Grouping and Camera Calibration in Complex Man-Made Environments
Grant Schindler and Frank Dellaert. We use the EM algorithm to perform a search over all continuous parameters that influence the location of a finite set of vanishing points in the scene.

ECCV 04: Multiview Reconstruction of Piecewise Smooth Subdivision Curves with a Variable Number of Control Points,
Michael Kaess, Rafal Zboinski, and Frank Dellaert. We investigate the automated reconstruction of piecewise smooth 3D curves, using tagged subdivision curves as a simple but flexible curve representation

ECCV 04: An MCMC-based Particle Filter for Tracking Multiple Interacting Targets
Zia Khan, Tucker Balch, Frank Dellaert. We describe a Markov chain Monte Carlo based particle filter that efficiently deals with multiple, interacting targets.

We also recently published a technical report on robust subspace modeling:


Robust Generative Subspace Modeling: The Subspace t Distribution
Z. Khan and F. Dellaert


Frank Dellaert Home
Computational Perception Lab

Frank Dellaert's Homepage