Skip to Main content Skip to Navigation
Theses

Déploiement sur une plate-forme de visualisation d'un algorithme coopératif pour la segmentation d'images IRM autour d'un système multi-agents

Abstract : The aim of this work is to deploy a segmentation system of magnetic resonance images (MRI) on a visualization platform "BrainVISA". The segmentation concern the classification of the brain into three regions : white matter, grey matter and cerebro-spinal fluid. There are several image segmentation algorithms ; each one with its advantages and its operational limits. In this work, we use two types of algorithms : The FCM (Fuzzy C-Mean) algorithm which uses the whole image and permits to manage imprecision and incertitude, and the region growing algorithm which takes account pixels neighbourhoods. The aim is to exploit the advantages of each method. The two segmentation methods are used in a cooperative way. The region growing predicate is adapted to our images in order to improve the segmented image quality. The implementation is then based on a multi-agents system (MAS) that allows to use efficiently the region growing algorithm. The segmentation is carried out on real images.
Document type :
Theses
Complete list of metadatas

Cited literature [19 references]  Display  Hide  Download

https://tel.archives-ouvertes.fr/tel-00568921
Contributor : Hadjer Laguel <>
Submitted on : Wednesday, February 23, 2011 - 9:03:54 PM
Last modification on : Friday, October 23, 2020 - 4:40:29 PM
Long-term archiving on: : Tuesday, November 6, 2012 - 2:50:57 PM

Identifiers

  • HAL Id : tel-00568921, version 1

Citation

Hadjer Laguel. Déploiement sur une plate-forme de visualisation d'un algorithme coopératif pour la segmentation d'images IRM autour d'un système multi-agents. Informatique [cs]. Université des Sciences et de la Technologie Houari Boumediene, 2010. Français. ⟨tel-00568921⟩

Share

Metrics

Record views

355

Files downloads

1943