Page d'accueil // SnT // News & E... // Research Seminar: Different paths to combining scene and camera self-calibration constraints

Research Seminar: Different paths to combining scene and camera self-calibration constraints

twitter linkedin google+ facebook email this page
Add to calendar
Conférencier : Dr. Adlane Habed, University of Bourgogne (France)
Date de l'événement : mercredi, 18 avril 2012, 15:30 - 16:15
Lieu : Campus Kirchberg – 6, rue Richard Coudenhove-Kalergi, Luxembourg, building F, room F213

Abstract: In a typical multi-view uncalibrated structure-from-motion pipeline, the three-dimensional projective structure of an unknown static scene is upgraded into a metric one through the self-calibration of the imaging camera. Camera self-calibration consists in retrieving the camera's intrinsic parameters solely using point correspondences across images and without resorting to the presence of a calibration pattern in the scene. However, this elegant and flexible approach turned out to be a difficult and challenging nonlinear problem exhibiting numerous degenerate, but common, camera motion sequences dooming self-calibration to failure. In order to cope with such issues, we have carried out several works to combine scene and/or camera knowledge with camera self-calibration constraints as to reduce the naturally existing ambiguities. This seminar covers some of the results we have obtained in this regard.

Adlane Habed received his Ph.D degree in Computer Science from the University of Sherbrooke (Qc., Canada) in 2005. He has served both as a Lecturer and an Assistant Professor at the School of Computer Science of the University of Windsor (On., Canada) between 2001 and 2007. In 2007, he joined the University of Bourgogne (France) and its CNRS-affiliated Electronics, Computer Science and Imaging Laboratory

(Le2i) as an Associate Professor. His research interests are mainly in Computer Vision and numerical optimization methods.