A Collaborative Software Tool for the Evaluation of MRI Brain Segmentation Methods - Université Clermont Auvergne
Communication Dans Un Congrès Année : 2007

A Collaborative Software Tool for the Evaluation of MRI Brain Segmentation Methods

Arvid Lundervold
  • Fonction : Auteur
  • PersonId : 1021630

Résumé

The paper presents two web services (AuroraWeb and AuroraWiki) for collaborative work on brain tissue seg-mentation and software tools related to MRI recordings. The web services allow members to share data and results by uploading and downloading, compare methods through a comprehensive segmentation evaluation framework, and provide overview and documentation of segmentation algorithms and their parameters. Presently the web services are applied to both simulated brain images (where a " ground truth " is available), and dual volume 3D brain images from a larger interdisciplinary study of cognitive aging. Commonly used brain segmenters that are tested include SPM2/SPM5, FSL and FreeSurfer, in addition to locally developed methods. New uploaded methods could either be as C/C++ source code, where a user template is provided, or as executables. We believe this technology can improve quality and dissemination of tools for quantitative MRI of the brain.
Fichier non déposé

Dates et versions

hal-01622924 , version 1 (24-10-2017)

Identifiants

Citer

Vincent Barra, Arvid Lundervold. A Collaborative Software Tool for the Evaluation of MRI Brain Segmentation Methods. International Symposium on Information Technology Convergence (ISITC 2007), Nov 2007, Jeonju, South Korea. pp.235 - 239, ⟨10.1109/ISITC.2007.9⟩. ⟨hal-01622924⟩
73 Consultations
0 Téléchargements

Altmetric

Partager

More