Deformable Organisms and Error Learning for Brain Segmentation

Abstract : Segmentation methods for medical images may not generalize well to different data sets or tasks, hampering their utility. We attempt to remedy these issues using deformable organisms to create an easily customizable segmentation plan. This plan is developed by borrowing ideas from artificial life to govern a set of deformable models that use control processes such as sensing, proactive planning, reactive behavior, and knowledge representation to segment an image. The image may have landmarks and features specific to that dataset; these may be easily incorporated into the plan. We validate this framework by creating a plan to locate the brain in 3D magnetic resonance images of the head (skull-stripping). This is important for surgical planning, understanding how diseases affect the brain, conducting longitudinal studies, registering brain data, and creating cortical surface models. Our plan dictates how deformable organisms find features in head images and cooperatively work to segment the brain. In addition, we use a method based on Adaboost to learn and correct errors in our segmentation. We tested our method on 630 T1-weighted images from healthy young adults, evaluating results using distance and overlap error metrics based on expert gold standard segmentations. We compare our segmentations with and without the error correction step; we also compare our results to three other widely used methods: BSE, BET, and the Hybrid Watershed algorithm. Our method had the least Hausdorff distance to expert segmentations on this dataset, but included slightly more non-brain voxels (false positives). Our framework captures diverse categories of information needed for skull-stripping, and produces competitive segmentations.
Type de document :
Communication dans un congrès
Pennec, Xavier and Joshi, Sarang and Nielsen, Mads. Proceedings of the Third International Workshop on Mathematical Foundations of Computational Anatomy - Geometrical and Statistical Methods for Modelling Biological Shape Variability, Sep 2011, Toronto, Canada. pp.135-147, 2011
Liste complète des métadonnées

Littérature citée [34 références]  Voir  Masquer  Télécharger

https://hal.inria.fr/inria-00624007
Contributeur : Xavier Pennec <>
Soumis le : jeudi 15 septembre 2011 - 15:34:00
Dernière modification le : mardi 16 janvier 2018 - 17:56:01
Document(s) archivé(s) le : vendredi 16 décembre 2011 - 02:25:40

Fichier

MFCA11_P_1.pdf
Fichiers produits par l'(les) auteur(s)

Identifiants

  • HAL Id : inria-00624007, version 1

Collections

Citation

Gautam Prasad, Anand A. Joshi, Albert Feng, Marina Barysheva, Katie L. Mcmahon, et al.. Deformable Organisms and Error Learning for Brain Segmentation. Pennec, Xavier and Joshi, Sarang and Nielsen, Mads. Proceedings of the Third International Workshop on Mathematical Foundations of Computational Anatomy - Geometrical and Statistical Methods for Modelling Biological Shape Variability, Sep 2011, Toronto, Canada. pp.135-147, 2011. 〈inria-00624007〉

Partager

Métriques

Consultations de la notice

551

Téléchargements de fichiers

612