Paper Info
Title
Fast deformable matching of 3D images over multiscale nested subspaces. Application to atlas-based MRI segmentation
Abstract
This paper presents a fast method to perform dense deformable matching of 3D images, applied to the registration of inter-subject brain MR images. To recover the complex morphological variations in neuroanatomy, the registration method uses a hierarchy of 3D deformations fields that are estimated, by minimizing a global energy function over a sequence of nested subspaces. The resulting deformable matching method shows low sensitivity to local minima and is able to track large non-linear deformations, with moderate computational load. The performances of the approach are assessed both on simulated 3D transformations and on a real data base of 3D brain MR images from different individuals. An application of the deformable image matching method to 3D atlas-based image segmentation is presented. This atlas-based segmentation is used at Strasbourg Hospital, in daily clinical applications, in order to extract regions of interest from 3D MR images of patients suffering from epilepsy.
Year
DOI
Venue
2003
10.1016/S0031-3203(02)00324-2
Pattern Recognition
Keywords
Field
DocType
3D hierarchical non-rigid image matching,Magnetic resonance imaging,3D brain atlas,Brain image segmentation
Computer vision,Pattern recognition,Image matching,Segmentation,Image segmentation,Linear subspace,Maxima and minima,Atlas (anatomy),Artificial intelligence,Mathematics
Journal
Volume
Issue
ISSN
36
8
0031-3203
Citations 
PageRank 
References 
8
0.84
32
Authors
3
Name
Order
Citations
PageRank
Olivier Musse1656.81
Fabrice Heitz240159.55
Jean-Paul Armspach322126.60