A general learning framework for non-rigid image registration

Guorong Wu, Feihu Qi, Dinggang Shen

Research output: Chapter in Book/Report/Conference proceedingConference contribution

7 Citations (Scopus)

Abstract

This paper presents a general learning framework for non-rigid registration of MR brain images. Given a set of training MR brain images, three major types of information are particularly learned, and further incorporated into a HAMMER registration algorithm for improving the performance of registration. First, the best features are learned from different types of local image descriptors for each part of brain, thereby the learned best features are consistent on the correspondence points across individual brains, but different on non-correspondence points. Moreover, the statistics of selected best features is learned from the training samples, and used to guide the feature matching during the image registration. Second, in order to avoid the local minima in the registration, the points hierarchically selected to drive image registration are determined by the learned consistency and distinctiveness of their respective best features. Third, deformation fields are adaptively represented by B-splines, with more control points placed on the regions with large shape variations across individual brains or on the regions with consistent and distinctive best features. Also, the statistics of B-splines based deformations is captured and used to regularize the brain registration. Finally, by incorporating all learned information into HAMMER registration framework, promising results are obtained on both real and simulated data.

Original languageEnglish
Title of host publicationLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Pages219-227
Number of pages9
Volume4091 LNCS
Publication statusPublished - 2006 Oct 10
Externally publishedYes
Event3rd International Workshop on Medical Imaging and Augmented Reality - Shanghai, China
Duration: 2006 Aug 172006 Aug 18

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume4091 LNCS
ISSN (Print)03029743
ISSN (Electronic)16113349

Other

Other3rd International Workshop on Medical Imaging and Augmented Reality
CountryChina
CityShanghai
Period06/8/1706/8/18

Fingerprint

Non-rigid Registration
Image registration
Image Registration
Brain
Registration
Learning
B-spline
Splines
Statistics
Feature Matching
Control Points
Training Samples
Local Minima
Descriptors
Framework
Correspondence

ASJC Scopus subject areas

  • Computer Science(all)
  • Biochemistry, Genetics and Molecular Biology(all)
  • Theoretical Computer Science

Cite this

Wu, G., Qi, F., & Shen, D. (2006). A general learning framework for non-rigid image registration. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 4091 LNCS, pp. 219-227). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 4091 LNCS).

A general learning framework for non-rigid image registration. / Wu, Guorong; Qi, Feihu; Shen, Dinggang.

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 4091 LNCS 2006. p. 219-227 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 4091 LNCS).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Wu, G, Qi, F & Shen, D 2006, A general learning framework for non-rigid image registration. in Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). vol. 4091 LNCS, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 4091 LNCS, pp. 219-227, 3rd International Workshop on Medical Imaging and Augmented Reality, Shanghai, China, 06/8/17.
Wu G, Qi F, Shen D. A general learning framework for non-rigid image registration. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 4091 LNCS. 2006. p. 219-227. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
Wu, Guorong ; Qi, Feihu ; Shen, Dinggang. / A general learning framework for non-rigid image registration. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 4091 LNCS 2006. pp. 219-227 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{c733e4c0fa874156b3b73bf4af8916d4,
title = "A general learning framework for non-rigid image registration",
abstract = "This paper presents a general learning framework for non-rigid registration of MR brain images. Given a set of training MR brain images, three major types of information are particularly learned, and further incorporated into a HAMMER registration algorithm for improving the performance of registration. First, the best features are learned from different types of local image descriptors for each part of brain, thereby the learned best features are consistent on the correspondence points across individual brains, but different on non-correspondence points. Moreover, the statistics of selected best features is learned from the training samples, and used to guide the feature matching during the image registration. Second, in order to avoid the local minima in the registration, the points hierarchically selected to drive image registration are determined by the learned consistency and distinctiveness of their respective best features. Third, deformation fields are adaptively represented by B-splines, with more control points placed on the regions with large shape variations across individual brains or on the regions with consistent and distinctive best features. Also, the statistics of B-splines based deformations is captured and used to regularize the brain registration. Finally, by incorporating all learned information into HAMMER registration framework, promising results are obtained on both real and simulated data.",
author = "Guorong Wu and Feihu Qi and Dinggang Shen",
year = "2006",
month = "10",
day = "10",
language = "English",
isbn = "3540372202",
volume = "4091 LNCS",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
pages = "219--227",
booktitle = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

}

TY - GEN

T1 - A general learning framework for non-rigid image registration

AU - Wu, Guorong

AU - Qi, Feihu

AU - Shen, Dinggang

PY - 2006/10/10

Y1 - 2006/10/10

N2 - This paper presents a general learning framework for non-rigid registration of MR brain images. Given a set of training MR brain images, three major types of information are particularly learned, and further incorporated into a HAMMER registration algorithm for improving the performance of registration. First, the best features are learned from different types of local image descriptors for each part of brain, thereby the learned best features are consistent on the correspondence points across individual brains, but different on non-correspondence points. Moreover, the statistics of selected best features is learned from the training samples, and used to guide the feature matching during the image registration. Second, in order to avoid the local minima in the registration, the points hierarchically selected to drive image registration are determined by the learned consistency and distinctiveness of their respective best features. Third, deformation fields are adaptively represented by B-splines, with more control points placed on the regions with large shape variations across individual brains or on the regions with consistent and distinctive best features. Also, the statistics of B-splines based deformations is captured and used to regularize the brain registration. Finally, by incorporating all learned information into HAMMER registration framework, promising results are obtained on both real and simulated data.

AB - This paper presents a general learning framework for non-rigid registration of MR brain images. Given a set of training MR brain images, three major types of information are particularly learned, and further incorporated into a HAMMER registration algorithm for improving the performance of registration. First, the best features are learned from different types of local image descriptors for each part of brain, thereby the learned best features are consistent on the correspondence points across individual brains, but different on non-correspondence points. Moreover, the statistics of selected best features is learned from the training samples, and used to guide the feature matching during the image registration. Second, in order to avoid the local minima in the registration, the points hierarchically selected to drive image registration are determined by the learned consistency and distinctiveness of their respective best features. Third, deformation fields are adaptively represented by B-splines, with more control points placed on the regions with large shape variations across individual brains or on the regions with consistent and distinctive best features. Also, the statistics of B-splines based deformations is captured and used to regularize the brain registration. Finally, by incorporating all learned information into HAMMER registration framework, promising results are obtained on both real and simulated data.

UR - http://www.scopus.com/inward/record.url?scp=33749394280&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=33749394280&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:33749394280

SN - 3540372202

SN - 9783540372202

VL - 4091 LNCS

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 219

EP - 227

BT - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

ER -