Topology-Preserving Registration: a Solution via Graph Cuts

Lucilio Cordero Grande; Gonzalo Vegas-Sánchez-Ferrero; Pablo Casaseca-de-la-Higuera; Carlos Alberola-López

Topology-Preserving Registration: a Solution via Graph Cuts

Lucilio Cordero Grande, Gonzalo Vegas-Sánchez-Ferrero, Pablo Casaseca-de-la-Higuera, Carlos Alberola-López

Universidad de Valladolid

Research output: Contribution to journal › Conference paper › peer-review

2 Citations (Scopus)

Abstract

In this work we propose a topology-preserving registration method based on a discrete Markov random field of deformations and a block-matching procedure. For that purpose, the fidelity of a given deformation to the data is established by a block-matching strategy, the smoothness of the transformation is favored by an appropriate prior on the field and topology preservation is guaranteed by imposing some hard-constraints on the local configurations of the field. The resulting deformation is defined as the maximum a posteriori of the field and it is estimated via graph cuts. Results on medical images show the efficiency of using graph cuts based fusion moves for the optimization of the field even though its potentials are neither sparse nor separable and the reduced fused problem turns to be non-submodular.

Original language	English
Pages (from-to)	420-431
Number of pages	11
Journal	Lecture Notes in Computer Science
Volume	6636
Publication status	Published - 2011

Keywords

image registration
topology preservation
Markov random fields
combinatorial optimization
graph cuts

Access to Document

http://link.springer.com/chapter/10.1007%2F978-3-642-21073-0_37

Cite this

@article{74ed96f8c76c479aa860003c19957ded,

title = "Topology-Preserving Registration: a Solution via Graph Cuts",

abstract = "In this work we propose a topology-preserving registration method based on a discrete Markov random field of deformations and a block-matching procedure. For that purpose, the fidelity of a given deformation to the data is established by a block-matching strategy, the smoothness of the transformation is favored by an appropriate prior on the field and topology preservation is guaranteed by imposing some hard-constraints on the local configurations of the field. The resulting deformation is defined as the maximum a posteriori of the field and it is estimated via graph cuts. Results on medical images show the efficiency of using graph cuts based fusion moves for the optimization of the field even though its potentials are neither sparse nor separable and the reduced fused problem turns to be non-submodular.",

keywords = "image registration, topology preservation, Markov random fields, combinatorial optimization, graph cuts",

author = "{Cordero Grande}, Lucilio and Gonzalo Vegas-S{\'a}nchez-Ferrero and Pablo Casaseca-de-la-Higuera and Carlos Alberola-L{\'o}pez",

year = "2011",

language = "English",

volume = "6636",

pages = "420--431",

journal = "Lecture Notes in Computer Science",

issn = "0302-9743",

publisher = "Springer Verlag",

}

TY - JOUR

T1 - Topology-Preserving Registration

T2 - a Solution via Graph Cuts

AU - Cordero Grande, Lucilio

AU - Vegas-Sánchez-Ferrero, Gonzalo

AU - Casaseca-de-la-Higuera, Pablo

AU - Alberola-López, Carlos

PY - 2011

Y1 - 2011

N2 - In this work we propose a topology-preserving registration method based on a discrete Markov random field of deformations and a block-matching procedure. For that purpose, the fidelity of a given deformation to the data is established by a block-matching strategy, the smoothness of the transformation is favored by an appropriate prior on the field and topology preservation is guaranteed by imposing some hard-constraints on the local configurations of the field. The resulting deformation is defined as the maximum a posteriori of the field and it is estimated via graph cuts. Results on medical images show the efficiency of using graph cuts based fusion moves for the optimization of the field even though its potentials are neither sparse nor separable and the reduced fused problem turns to be non-submodular.

AB - In this work we propose a topology-preserving registration method based on a discrete Markov random field of deformations and a block-matching procedure. For that purpose, the fidelity of a given deformation to the data is established by a block-matching strategy, the smoothness of the transformation is favored by an appropriate prior on the field and topology preservation is guaranteed by imposing some hard-constraints on the local configurations of the field. The resulting deformation is defined as the maximum a posteriori of the field and it is estimated via graph cuts. Results on medical images show the efficiency of using graph cuts based fusion moves for the optimization of the field even though its potentials are neither sparse nor separable and the reduced fused problem turns to be non-submodular.

KW - image registration

KW - topology preservation

KW - Markov random fields

KW - combinatorial optimization

KW - graph cuts

M3 - Conference paper

SN - 0302-9743

VL - 6636

SP - 420

EP - 431

JO - Lecture Notes in Computer Science

JF - Lecture Notes in Computer Science

ER -

Topology-Preserving Registration: a Solution via Graph Cuts

Abstract

Keywords

Access to Document

Fingerprint

Cite this