A general fast registration framework by learning deformation-appearance correlation

Minjeong Kim, Guorong Wu, Pew Thian Yap, Dinggang Shen

Research output: Contribution to journalArticle

22 Citations (Scopus)

Abstract

In this paper, we propose a general framework for performance improvement of the current state-of-the-art registration algorithms in terms of both accuracy and computation time. The key concept involves rapid prediction of a deformation field for registration initialization, which is achieved by a statistical correlation model learned between image appearances and deformation fields. This allows us to immediately bring a template image as close as possible to a subject image that we need to register. The task of the registration algorithm is hence reduced to estimating small deformation between the subject image and the initially warped template image, i.e., the intermediate template (IT). Specifically, to obtain a good subject-specific initial deformation, support vector regression is utilized to determine the correlation between image appearances and their respective deformation fields. When registering a new subject onto the template, an initial deformation field is first predicted based on the subject's image appearance for generating an IT. With the IT, only the residual deformation needs to be estimated, presenting much less challenge to the existing registration algorithms. Our learning-based framework affords two important advantages: 1) by requiring only the estimation of the residual deformation between the IT and the subject image, the computation time can be greatly reduced; 2) by leveraging good deformation initialization, local minima giving suboptimal solution could be avoided. Our framework has been extensively evaluated using medical images from different sources, and the results indicate that, on top of accuracy improvement, significant registration speedup can be achieved, as compared with the case where no prediction of initial deformation is performed.

Original languageEnglish
Article number6035779
Pages (from-to)1823-1833
Number of pages11
JournalIEEE Transactions on Image Processing
Volume21
Issue number4
DOIs
Publication statusPublished - 2012 Apr 1
Externally publishedYes

Fingerprint

Learning
Statistical Models

Keywords

  • Deformation prediction
  • fast image registration
  • principal component analysis (PCA)
  • support vector regression (SVR)

ASJC Scopus subject areas

  • Computer Graphics and Computer-Aided Design
  • Software
  • Medicine(all)

Cite this

A general fast registration framework by learning deformation-appearance correlation. / Kim, Minjeong; Wu, Guorong; Yap, Pew Thian; Shen, Dinggang.

In: IEEE Transactions on Image Processing, Vol. 21, No. 4, 6035779, 01.04.2012, p. 1823-1833.

Research output: Contribution to journalArticle

Kim, Minjeong ; Wu, Guorong ; Yap, Pew Thian ; Shen, Dinggang. / A general fast registration framework by learning deformation-appearance correlation. In: IEEE Transactions on Image Processing. 2012 ; Vol. 21, No. 4. pp. 1823-1833.
@article{9ccd5cbfac4e45fa8f326bc503f8f74b,
title = "A general fast registration framework by learning deformation-appearance correlation",
abstract = "In this paper, we propose a general framework for performance improvement of the current state-of-the-art registration algorithms in terms of both accuracy and computation time. The key concept involves rapid prediction of a deformation field for registration initialization, which is achieved by a statistical correlation model learned between image appearances and deformation fields. This allows us to immediately bring a template image as close as possible to a subject image that we need to register. The task of the registration algorithm is hence reduced to estimating small deformation between the subject image and the initially warped template image, i.e., the intermediate template (IT). Specifically, to obtain a good subject-specific initial deformation, support vector regression is utilized to determine the correlation between image appearances and their respective deformation fields. When registering a new subject onto the template, an initial deformation field is first predicted based on the subject's image appearance for generating an IT. With the IT, only the residual deformation needs to be estimated, presenting much less challenge to the existing registration algorithms. Our learning-based framework affords two important advantages: 1) by requiring only the estimation of the residual deformation between the IT and the subject image, the computation time can be greatly reduced; 2) by leveraging good deformation initialization, local minima giving suboptimal solution could be avoided. Our framework has been extensively evaluated using medical images from different sources, and the results indicate that, on top of accuracy improvement, significant registration speedup can be achieved, as compared with the case where no prediction of initial deformation is performed.",
keywords = "Deformation prediction, fast image registration, principal component analysis (PCA), support vector regression (SVR)",
author = "Minjeong Kim and Guorong Wu and Yap, {Pew Thian} and Dinggang Shen",
year = "2012",
month = "4",
day = "1",
doi = "10.1109/TIP.2011.2170698",
language = "English",
volume = "21",
pages = "1823--1833",
journal = "IEEE Transactions on Image Processing",
issn = "1057-7149",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
number = "4",

}

TY - JOUR

T1 - A general fast registration framework by learning deformation-appearance correlation

AU - Kim, Minjeong

AU - Wu, Guorong

AU - Yap, Pew Thian

AU - Shen, Dinggang

PY - 2012/4/1

Y1 - 2012/4/1

N2 - In this paper, we propose a general framework for performance improvement of the current state-of-the-art registration algorithms in terms of both accuracy and computation time. The key concept involves rapid prediction of a deformation field for registration initialization, which is achieved by a statistical correlation model learned between image appearances and deformation fields. This allows us to immediately bring a template image as close as possible to a subject image that we need to register. The task of the registration algorithm is hence reduced to estimating small deformation between the subject image and the initially warped template image, i.e., the intermediate template (IT). Specifically, to obtain a good subject-specific initial deformation, support vector regression is utilized to determine the correlation between image appearances and their respective deformation fields. When registering a new subject onto the template, an initial deformation field is first predicted based on the subject's image appearance for generating an IT. With the IT, only the residual deformation needs to be estimated, presenting much less challenge to the existing registration algorithms. Our learning-based framework affords two important advantages: 1) by requiring only the estimation of the residual deformation between the IT and the subject image, the computation time can be greatly reduced; 2) by leveraging good deformation initialization, local minima giving suboptimal solution could be avoided. Our framework has been extensively evaluated using medical images from different sources, and the results indicate that, on top of accuracy improvement, significant registration speedup can be achieved, as compared with the case where no prediction of initial deformation is performed.

AB - In this paper, we propose a general framework for performance improvement of the current state-of-the-art registration algorithms in terms of both accuracy and computation time. The key concept involves rapid prediction of a deformation field for registration initialization, which is achieved by a statistical correlation model learned between image appearances and deformation fields. This allows us to immediately bring a template image as close as possible to a subject image that we need to register. The task of the registration algorithm is hence reduced to estimating small deformation between the subject image and the initially warped template image, i.e., the intermediate template (IT). Specifically, to obtain a good subject-specific initial deformation, support vector regression is utilized to determine the correlation between image appearances and their respective deformation fields. When registering a new subject onto the template, an initial deformation field is first predicted based on the subject's image appearance for generating an IT. With the IT, only the residual deformation needs to be estimated, presenting much less challenge to the existing registration algorithms. Our learning-based framework affords two important advantages: 1) by requiring only the estimation of the residual deformation between the IT and the subject image, the computation time can be greatly reduced; 2) by leveraging good deformation initialization, local minima giving suboptimal solution could be avoided. Our framework has been extensively evaluated using medical images from different sources, and the results indicate that, on top of accuracy improvement, significant registration speedup can be achieved, as compared with the case where no prediction of initial deformation is performed.

KW - Deformation prediction

KW - fast image registration

KW - principal component analysis (PCA)

KW - support vector regression (SVR)

UR - http://www.scopus.com/inward/record.url?scp=84859010761&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84859010761&partnerID=8YFLogxK

U2 - 10.1109/TIP.2011.2170698

DO - 10.1109/TIP.2011.2170698

M3 - Article

C2 - 21984505

AN - SCOPUS:84859010761

VL - 21

SP - 1823

EP - 1833

JO - IEEE Transactions on Image Processing

JF - IEEE Transactions on Image Processing

SN - 1057-7149

IS - 4

M1 - 6035779

ER -