Department of Computer Science
University of North Carolina at Chapel Hill
Visual 3D Modeling Using Cameras and Camera Networks
In this talk we will discuss how 3D models can be recovered from images recorded with one or more cameras. Our focus consist of developing more flexible approaches with fewer limitations. Our prior work in this area allows to obtain detailed 3D models from images acquired with an uncalibrated hand-held camera. For the single camera case, we will discuss recent results on radiometric auto-calibration and on linear calibration of omnidirectional cameras. We will also discuss a factorization approach that can deal with articulated motions.
Next, the calibration and synchronization of camera networks is addressed. An efficient and robust algorithm that computes the epipolar geometry from silhouttes of dynamic objects is presented. This makes this approach particularly suitable for visual hull systems. Using the pairwise epipolar geometries, the complete calibration of the network is computed, and refined through bundle adjustment. This approach is extended to deal with unsynchronized video streams. We discuss the impact of having unsynchronized video streams on 3D reconstruction, discuss possibilities to optimally combine silhouette and photo-consistency constraints and address the reconstruction of specular surfaces. Ongoing efforts to extend this research to active pan-tilt-zoom camera networks are also briefly discussed.
Marc Pollefeys obtained his masters in Electrical Engineering and his PhD from the Katholieke Universiteit Leuven in 1994 and 1999 respectively. He is currently an assistant professor in the Computer Science department of the University of North Carolina at Chapel Hill.
His main area of research is computer vision. One of his main research goals is to develop flexible approaches to capture visual representations of real world objects, scenes and events. Dr. Pollefeys has received several prizes and awards for his research, including the prestigious Marr prize at ICCV '98 and an NSF CAREER award. He is the author or co-author of more than 90 technical papers on structure from motion, self-calibration, multiple-view geometry, stereo, 3D modeling, camera networks, omnidirectional vision, projector-camera systems, image-based rendering and applications. He has organized workshops and courses at major vision and graphics conferences and has served on the program committees of many conferences. He is a regular reviewer for most of the major vision, graphics and photogrammetry journals.
Dr. Pollefeys is a member of the editorial board of Foundations and Trends in Computer Graphics and Computer Vision. His research is funded by NSF, Darpa, Siemens and the State of North Carolina.
Hosts: Rich Radke (ECSE), x6483
Chuck Stewart (CS), x6731
Wednesday, February 16, 2005
JEC 3117 - 2:30 p.m. to 3:30 p.m.
Refreshments at 2:00 p.m.