|
BMVC 2011
The 22nd British Machine Vision Conference
University of Dundee, 29 August - 2 September 2011
|
Keynote Speakers
Tracking and Understanding Human Motion
The estimation and interpretation of human motion from video are key enabling technologies for myriad applications. But despite more than a decade of focused research, the general problem in natural environments, from monocular observations, remains challenging. I will outline the nature of the problem, and describe some recent advances in modeling human motion, 3D pose estimation, the inference of human attributes, and on the inference of peoples' interactions with their environments, based on principles of kinematics, biomechanics and mechanics.
|
|
David Fleet is Professor of Computer Science at the University of Toronto. He received the PhD in Computer Science in 1991, after which he was on faculty at Queen's University, and then with the Palo Alto Research Center (PARC). His main research interests and publications span the areas of computer vision, image processing, visual perception, and visual neuroscience. In 1996 he was awarded an Alfred P. Sloan Research Fellowship for his research on biological vision. He has won paper awards at ICCV 1999, CVPR 2001, UIST 2003, and BMVC 2009. In 2010, he was awarded the Koenderink Prize for his work with Michael Black and Hedvig Sidenbladh on human pose tracking. He has served as Area Chair for numerous computer vision and machine learning conferences. He was Program Co-Chair for CVPR 2003, and will be Program Co-Chair for ECCV 2014. He has been Associate Editor (2000-2004), Associate Editor-In-Chief (2004-2008), and is currently on the Advisory Board for IEEE Transactions on PAMI. He is Fellow of the Canadian Institute for Advanced Research. |
Shape Grammars and Procedural Modeling towards Large Scale 3D Modeling and Reconstruction
Large scale realistic urban modeling has became a core component for a number of domains like the games industry, urban planning, etc. In such a context, scalability, modularity, compactness and context are to be reconciled. In this talk, we adopt shape grammars, a powerful structural representation that naturally encompasses the aforementioned objectives, and introduce novel inference methods to address large scale image-based modeling. We adopt reinforcement learning for facade parsing and evolutionary optimization methods towards addressing complete 3D reconstruction through fusion of visual and depth data. Promising results demonstrate the potential of our methods.
Joint Work with: L. Simon, O. Teboul & P. Koutsourakis
|
|
Nikos Paragios is professor of Applied Mathematics and Computer Science, director of the Center for Visual Computing of Ècole Centrale de Paris & Ècole des Ponts - ParisTech, and scientific leader of GALEN group of Ècole Centrale de Paris / INRIA Saclay, Ile-de-France. Prof. Paragios is an IEEE fellow, participates in numerous editorial boards (IEEE T-PAMI, IJCV, CVIU MedIA,), has published more than 150 papers in top conferences and journals of his field and has also 17 international granted patents. His interests are in the areas of computer vision, medical imaging, remote sensing and human computer interaction. B.Sc. (highest honors, valedictorian) and M.Sc. (highest honors) in Computer Science [University of Crete (Greece) - 1994,1996] , Ph.D. (highest honors) in electrical and computer engineering [I.N.R.I.A., 2000] HDR (Habilitation a Diriger de Recherches - D.Sc.) [University of Nice/Sophia Antipolis (France), 2005)]. |
|
Last modified on 15 July 2011, 11.05am
Design based on Liblueen by Nicolas Fafchamps Site maintained by Stephen McKenna |