Ομιλία Καθ. Ιάσονα Κόκκινου, 09/06/2009 PDF Print E-mail
Saturday, 06 June 2009 11:17

Το φοιτητικό παράρτημα της ΙΕΕΕ του ΕΜΠ σας προσκαλεί στην ομιλία του

Καθηγητή Ιάσονα Κόκκινου

Ecole Centrale de Paris/INRIA-Saclay

Object Parsing using Hierarchical Compositional Models

Τρίτη 9 Ιουνίου 2009, 11:30

Αμφιθέατρο Πολυμέσων - Κεντρική Βιβλιοθήκη ΕΜΠ

Θα προηγηθεί ομιλία στα πλαίσια του μαθήματος της Όρασης Υπολογιστών,
η οποία θα ξεκινήσει στις 10:30 στον ίδιο χώρο

Περίληψη ομιλίας:

The ultimate goal of high-level vision is to parse images, namely identify the objects inside an image, delineate their borders, and recover their pose and decomposition into simpler parts. This naturally leads to hierarchical object representations that recursively decompose an object into simpler structures, in the same way that a grammar decomposes a sentence into its constituent phrases. The parsing problem then consists in efficiently finding an optimal way of putting together `visual words', i.e. the simplest image structures in order to build an object.

To solve this problem, we will develop an approach that guides search by combining bottom-up (image-based) and top-down (model-based) information based on the A* algorithm, thereby resulting in a tenfold speedup when compared to purely bottom-up search. This approach is shown to simultaneously localize and parse objects in natural images containing heavy clutter and is experimentally validated on challenging object detection benchmarks.


Σύντομο βιογραφικό:

Iasonas Kokkinos obtained his Diploma and Ph.D. degrees in Electrical and Computer Engineering from the National Technical University of Athens in 2001 and 2006 respectively. During 2006-2008 he was a postdoctoral scholar at the University of California at Los Angeles and is currently an assistant professor at Ecole Centrale de Paris and affiliate researcher at INRIA-Saclay. He has worked on texture analysis and segmentation, object detection and deformable models, while his current research focus is on jointly solving low- and high- level computer vision problems in a common, probabilistic framework. He has obtained several awards for academic excellence, including the Paris Kanellakis award, and scholarships from the Bodossaki foundation and the National Scholarship Foundation. His work has appeared in the most selective conferences and academic journals, while he has served on the program committee of the leading computer vision conferences.
URL: http://www.mas.ecp.fr/vision/Personnel/iasonas/

Πληροφορίες: ΙΕΕΕ NTUA Student Branch, ieeesb[at]cslab.ece.ntua.gr

Last Updated ( Wednesday, 10 June 2009 12:41 )
 
Valid XHTML & CSS | Template Design ah-68 | Copyright © 2009 by Firma