Active shape models—their training and application
Computer Vision and Image Understanding
Diffeomorphisms Groups and Pattern Matching in Image Analysis
International Journal of Computer Vision
Variational problems on flows of diffeomorphisms for image matching
Quarterly of Applied Mathematics
Geodesic Shooting for Computational Anatomy
Journal of Mathematical Imaging and Vision
Landmark matching via large deformation diffeomorphisms
IEEE Transactions on Image Processing
On the Manifold Structure of the Space of Brain Images
MICCAI '09 Proceedings of the 12th International Conference on Medical Image Computing and Computer-Assisted Intervention: Part I
Atlas construction and image analysis using statistical cardiac models
STACOM'10/CESC'10 Proceedings of the First international conference on Statistical atlases and computational models of the heart, and international conference on Cardiac electrophysiological simulation challenge
On the Geometry of Multivariate Generalized Gaussian Models
Journal of Mathematical Imaging and Vision
| Hi-index | 7.29 |
This paper considers the problem of defining a parameterization (chart) on the group of diffeomorphisms with compact support, motivated primarily by a problem in image registration, where diffeomorphic warps are used to align images. Constructing a chart on the diffeomorphism group will enable the quantitative analysis of these warps to discover the normal and abnormal variation of structures in a population. We construct a chart for particular choices of boundary conditions on the space on which the diffeomorphism acts, and for a particular class of metrics on the diffeomorphism group, which define a class of diffeomorphic interpolating splines. The geodesic equation is computed for this class of metrics, and we show how it can be solved in the spline representation. Furthermore, we demonstrate that the spline representation generates submanifolds of the diffeomorphism group, and we study this mapping. Explicit computational examples are included, showing how this chart can be constructed in practice, and that the use of the geodesic distance allows better classification of variation than those obtained using just a Euclidean metric on the space of warps.