markovian Image Models: Applications in Unsupervised Image Segmentation  Page description

Help  Print 
Back »

 

Details of project

 
Identifier
46805
Type K
Principal investigator Kató, Zoltán
Title in Hungarian Markov mezők a képmodellezésben, alkalmazásuk az automatikus képszegmentálás területén.
Title in English markovian Image Models: Applications in Unsupervised Image Segmentation
Panel Informatics and Electrical Engineering
Department or equivalent Department of Image Processing and Computer Graphics (University of Szeged)
Starting date 2004-01-01
Closing date 2006-12-31
Funding (in million HUF) 2.018
FTE (full time equivalent) 0.00
state closed project





 

Final report

 
Results in Hungarian
1) Kifejlesztettünk egy olyan szín és textúra alapú szegmentáló MRF algoritmust, amely alkalmas egy kép automatikus szegmentálását elvégezni. Az eredményeinket az Image and Vision Computing folyóiratban publikáltuk. 2) Kifejlesztettünk egy Reversible Jump Markov Chain Monte Carlo technikán alapuló automatikus képszegmentáló eljárást, melyet sikeresen alkalmaztunk színes képek teljesen automatikus szegmentálására. Az eredményeinket a BMVC 2004 konferencián és az Image and Vision Computing folyóiratban publikáltuk. 3) A modell többrétegű továbbfejlesztését alkalmaztuk video objektumok szín és mozgás alapú szegmentálására, melynek eredményeit a HACIPPR 2005 illetve az ACCV 2006 nemzetközi konferenciákon publikáltuk. Szintén ehhez az alapproblémához kapcsolódik Horváth Péter hallgatómmal az optic flow szamításával illetve szín, textúra és mozgás alapú GVF aktív kontúrral kapcsoltos munkáink. TDK dolgozata első helyezést ért el a 2004-es helyi versenyen, az eredményeinket pedig a KEPAF 2004 konferencián publikáltuk. 4) Horváth Péter PhD hallgatómmal illetve az franciaországi INRIA Ariana csoportjával, kidolgoztunk egy olyan képszegmentáló eljárást, amely a szegmentálandó objektum alakját is figyelembe veszi. Az eredményeinket az ICPR 2006 illetve az ICCVGIP 2006 konferencián foglaltuk össze. A modell előzményeként kidolgoztunk továbbá egy alakzat-momemntumokon alapuló aktív kontúr modellt, amelyet a HACIPPR 2005 konferencián publikáltunk.
Results in English
1) We have proposed a monogrid MRF model which is able to combine color and texture features in order to improve the quality of segmentation results. We have also solved the estimation of model parameters. This work has been published in the Image and Vision Computing journal. 2) We have proposed an RJMCMC sampling method which is able to identify multi-dimensional Gaussian mixtures. Using this technique, we have developed a fully automatic color image segmentation algorithm. Our results have been published at BMVC 2004 international conference and in the Image and Vision Computing journal. 3) A new multilayer MRF model has been proposed which is able to segment an image based on multiple cues (such as color, texture, or motion). This work has been published at HACIPPR 2005 and ACCV 2006 international conferences. The work on optic flow computation and color-, texture-, and motion-based GVF active contours doen with my student, Mr. Peter Horvath, won a first price at the local Student Research Competition in 2004. Results have been presented at KEPAF 2004 conference. 4) A new shape prior, called 'gas of circles' has been introduced using active contour models. This work is done in collaboration with the Ariana group of INRIA, France and my PhD student, Mr. Peter Horvath. Results are published at the ICPR 2006 and ICCVGIP 2006 conferences. A preliminary study on active contour models using shape-moments has also been done, these results are published at HACIPPR 2005.
Full text http://real.mtak.hu/1584/
Decision
Yes





 

List of publications

 
Zoltan Kato and Ting Chuen Pong: A Multi-Layer MRF Model for Video Object Segmentation, P. J. Narayanan, Shree K. Nayar, and Heung-Yeung Shum, editors, Asian Conference on Computer Vision, vol. 3852 of LNCS, Hyderabad, India, pp. 953-962, Springer, 2006
Zoltan Kato: Reversible Jump Markov Chain Monte Carlo for Unsupervised MRF Color Image Segmentation, Proceedings of Brithish Machine Vision Conference, volume 1, Kingston, UK, pp 37-46, September, 2004
Zoltan Kato and Ting Chuen Pong: Video Object Segmentation Using a Multicue Markovian Model, Proceedings of Hungarian-Austrian Conference on Image Processing and Pattern Recognition, Veszprem, Hungary, pp 111-118, May, 2005
Peter Horvath, Avik Bhattacharya, Ian H. Jermyn, Josiane Zerubia and Zoltan Kato: Shape Moments for Region-Based Active Contours, Proceedings of Hungarian-Austrian Conference on Image Processing and Pattern Recognition, Veszprem, Hungary, pp 187-194, May, 2005
Zoltan Kato: Segmentation of Color Images via Reversible Jump MCMC Sampling, Image and Vision Computing, Elsevier (accepted, in press, available at http://dx.doi.org/10.1016/j.imavis.2006.12.004), 2007
Zoltan Kato and Ting Chuen Pong: A Markov Random Field Image Segmentation Model for Color Textured Images, Image and Vision Computing, vol. 24 no. 10, pp 1103-1114, Elsevier, 2006
Peter Horvath, Ian Jermyn, Zoltan Kato, and Josiane Zerubia: An Improved `Gas of Circles' Higher-Order Active Contour Model and its Application to Tree Crown Extraction, Prem Kalra and Shmuel Peleg, editors, Proc. Indian Conference on Computer Vision, Graphics and Image Processing, vol. 4338 of LNCS, Madurai, India, pp 152-161, Springer, 2006
Peter Horvath, Ian Jermyn, Zoltan Kato, and Josiane Zerubia: A Higher-Order Active Contour Model for Tree Detection, Proceedings of International Conference on Pattern Recognition, Hong Kong, China, pp 130-133, August, IAPR, IEEE, 2006




Back »