Login
Talks & Seminars
Energy Minimization Methods in Image Segmentation
Zoltan Kato, Institute of Informatics, University of Szeged, Hungary
Date & Time: January 24, 2006 14:00
Venue: EE Conference Room
Abstract:
In this talk, we present some of our recent research works on image segmentation using probabilistic and variational approaches. The talk consists of 3 main parts:
  1. After a brief introduction to MRF modelization, we present a multi-layer probabilistic model which efficiently integrates different visual cues (like color, texture or motion) for segmentation.
  2. Probabilistic modelization usually requires some parameter tuning which can be solved using standard statistical approaches. However, model order selection is rararely addressed due to the inherent difficulties in simultanously estimating the model parameters and generating samples from the underlying posterior distribution. We adopt Reversible Jump MCMC to solve this problem.
  3. Active Contours are well known for producing closed contours, where the shape of the contour is controlled by the internal forces. In this part, we will present two shape priors (one is a template-like prior and the other one is a generic structural prior) which is then used in a variational segmentation framework in order to deal with high noise, occlusions, and cluttered background.
Speaker Profile:
Zoltan Kato received the BS and MS degrees in computer science from the Jozsef Attila University, Szeged, Hungary in 1988 and 1990, and the PhD degree from University of Nice doing his research at INRIA -- Sophia Antipolis, France in 1994. Since then, he has been a visiting research associate at the Computer Science Department of the Hong Kong University of Science & Technology, Hong Kong; an ERCIM postdoc fellow at CWI, Amsterdam, The Netherlands; and a visiting fellow at the School of Computing, National University of Singapore. In 2002, he joined the Institute of Informatics, University of Szeged, Hungary. His research interests include image segmentation, statistical image models, Markov random fields, color, texture, motion, content based retrieval, combinatorial optimization, parameter estimation, shape modeling, variational and level set methods. He is currently an Associate Editor of IEEE Transactions on Image Processing, Managing Editor of Acta Cybernetica, member of the IEEE Signal Processing Society, and British Machine Vision Association.

http://www.inf.u-szeged.hu/~kato/
List of Talks

Webmail

Username:
Password:
Faculty CSE IT
Forgot Password
    [+] Sitemap     Feedback