SEGMENTATION D’IMAGES TYPE RMN FIXE PAR LES CHAINES DE MARKOV CACHEES, APRENTISSAGE NON SUPPERVISE ET MINIISATION DE LA FONCTION D’ENERGIE

RECHID, NAIMA (2004) SEGMENTATION D’IMAGES TYPE RMN FIXE PAR LES CHAINES DE MARKOV CACHEES, APRENTISSAGE NON SUPPERVISE ET MINIISATION DE LA FONCTION D’ENERGIE. Masters thesis, Université Mohamed Khider - Biskra.

[img]
Preview
Text
PAGE DE GARDE.pdf

Download (20kB) | Preview

Abstract

The most appropriate system to observe subject anatomy is nuclear magnetic resonance imaging (NMR), and a major issue of image processing is to segment automatically anatomy structures. This is the scope of our work. Our contribution has been to present an unsupervised segmentation method which the formalism is relied on hidden Markov field and chain theory. The originality of this method is, it takes into account structural information processed as flexible spatial. Our first results are very satisfactory. One can extend our method of segmentation to the other types of images.

Item Type: Thesis (Masters)
Uncontrolled Keywords: Hidden Markov field, Hidden Markov chain, Gibbs sampler, Algorithm ICE, Algorithm MPM, unsupervised segmentation, Image NMR.
Subjects: T Technology > TK Electrical engineering. Electronics Nuclear engineering
Divisions: Faculté des Sciences et de la technologie > Département de Génie Electrique
Depositing User: Admin02 TMLBiskra
Date Deposited: 08 Nov 2014 17:39
Last Modified: 08 Nov 2014 17:39
URI: http://thesis.univ-biskra.dz/id/eprint/689

Actions (login required)

View Item View Item