Current page: Members->List of Members->Vaclav Hlavac
 

Vaclav Hlavac
Czech Technical University
Membership Number: 39
Address: Centre for Machine Perception, Department of Cybernetics, 121 35 Prague 2, Karlovo namesti 13, Czech Republic
Email: hlavac@fel.cvut.cz
Phone: +420 2 24357465
Fax: +420 2 24357385
URL: http://cmp.felk.cvut.cz

Biographical Sketch
Prof. Vaclav Hlavac (1956) joined the Czech Technical University in 1982 as a PhD student. Currently heads the Center for Machine Perception, Department of Cybernetics, Faculty of Electrical Engineering, Czech Technical University in Prague. He has been appointed full professor in 1998. His research interests are in 3D computer vision, statistical and structural pattern recognition. He has co-authored over 100 of papers and three books. He is a Principal Investigator in Prague for IST projects ActIPret and ISAAC.

Centre for Machine Perception
The Center for Machine Perception (abbreviated CMP, established in 1996) is a research and teaching unit active in the fields of computer vision, pattern recognition, and mathematical modelling of uncertainty. CMP is involved in the EU project IST-2001-32184 ActIPret (Interpreting and Understanding Activities of Expert Operators for Teaching and Education).

Knowledge Representation: We are interested in relation between structural representation (e.g., by a formal language) and statistical representation. Our computer vision research touches representation issues too. Probably the strongest is the 3D reconstruction.

Learning: (1) We are interested in statistical pattern recognition approach to learning. Recently we contributed to learning support vector machines and its computationally more efficient alternative. We also studied convergence properties of the EM algorithm and find a niche where it converges globally. (2) We are interested in transition between statistical and structural pattern recognition including learning issues. For our standpoints see our recent book: Schlesinger M.I., Hlavac V.: Ten lectures on statistical and structural pattern recognition, Kluwer, Dordrecht, The Netherlands, May 2002.

Recognition and Categorization: We are active in statistical pattern recognition that covers several recognition issues. Also, the affine invariant frames suggested recently by J. Matas contribute to image-based recognition (image database retrieval, recognition for ActIPret project).


Site generated on Friday, 06 January 2006