Predict brain MR image registration via sparse learning of appearance and transformation

Qian Wang, Minjeong Kim, Yonghong Shi, Guorong Wu, Dinggang Shen

Research output: Contribution to journalArticle

24 Citations (Scopus)

Abstract

We propose a new approach to register the subject image with the template by leveraging a set of intermediate images that are pre-aligned to the template. We argue that, if points in the subject and the intermediate images share similar local appearances, they may have common correspondence in the template. In this way, we learn the sparse representation of a certain subject point to reveal several similar candidate points in the intermediate images. Each selected intermediate candidate can bridge the correspondence from the subject point to the template space, thus predicting the transformation associated with the subject point at the confidence level that relates to the learned sparse coefficient. Following this strategy, we first predict transformations at selected key points, and retain multiple predictions on each key point, instead of allowing only a single correspondence. Then, by utilizing all key points and their predictions with varying confidences, we adaptively reconstruct the dense transformation field that warps the subject to the template. We further embed the prediction-reconstruction protocol above into a multi-resolution hierarchy. In the final, we refine our estimated transformation field via existing registration method in effective manners. We apply our method to registering brain MR images, and conclude that the proposed framework is competent to improve registration performances substantially.

Original languageEnglish
Pages (from-to)61-75
Number of pages15
JournalMedical Image Analysis
Volume20
Issue number1
DOIs
Publication statusPublished - 2015 Feb 1

Fingerprint

Image registration
Brain
Learning

Keywords

  • Brain MR image registration
  • Correspondence detection
  • Deformable image registration
  • Sparsity learning
  • Transformation prediction

ASJC Scopus subject areas

  • Computer Graphics and Computer-Aided Design
  • Computer Vision and Pattern Recognition
  • Radiology Nuclear Medicine and imaging
  • Health Informatics
  • Radiological and Ultrasound Technology

Cite this

Predict brain MR image registration via sparse learning of appearance and transformation. / Wang, Qian; Kim, Minjeong; Shi, Yonghong; Wu, Guorong; Shen, Dinggang.

In: Medical Image Analysis, Vol. 20, No. 1, 01.02.2015, p. 61-75.

Research output: Contribution to journalArticle

Wang, Qian ; Kim, Minjeong ; Shi, Yonghong ; Wu, Guorong ; Shen, Dinggang. / Predict brain MR image registration via sparse learning of appearance and transformation. In: Medical Image Analysis. 2015 ; Vol. 20, No. 1. pp. 61-75.
@article{f4b0045f175849b4b6930201e925d1ea,
title = "Predict brain MR image registration via sparse learning of appearance and transformation",
abstract = "We propose a new approach to register the subject image with the template by leveraging a set of intermediate images that are pre-aligned to the template. We argue that, if points in the subject and the intermediate images share similar local appearances, they may have common correspondence in the template. In this way, we learn the sparse representation of a certain subject point to reveal several similar candidate points in the intermediate images. Each selected intermediate candidate can bridge the correspondence from the subject point to the template space, thus predicting the transformation associated with the subject point at the confidence level that relates to the learned sparse coefficient. Following this strategy, we first predict transformations at selected key points, and retain multiple predictions on each key point, instead of allowing only a single correspondence. Then, by utilizing all key points and their predictions with varying confidences, we adaptively reconstruct the dense transformation field that warps the subject to the template. We further embed the prediction-reconstruction protocol above into a multi-resolution hierarchy. In the final, we refine our estimated transformation field via existing registration method in effective manners. We apply our method to registering brain MR images, and conclude that the proposed framework is competent to improve registration performances substantially.",
keywords = "Brain MR image registration, Correspondence detection, Deformable image registration, Sparsity learning, Transformation prediction",
author = "Qian Wang and Minjeong Kim and Yonghong Shi and Guorong Wu and Dinggang Shen",
year = "2015",
month = "2",
day = "1",
doi = "10.1016/j.media.2014.10.007",
language = "English",
volume = "20",
pages = "61--75",
journal = "Medical Image Analysis",
issn = "1361-8415",
publisher = "Elsevier",
number = "1",

}

TY - JOUR

T1 - Predict brain MR image registration via sparse learning of appearance and transformation

AU - Wang, Qian

AU - Kim, Minjeong

AU - Shi, Yonghong

AU - Wu, Guorong

AU - Shen, Dinggang

PY - 2015/2/1

Y1 - 2015/2/1

N2 - We propose a new approach to register the subject image with the template by leveraging a set of intermediate images that are pre-aligned to the template. We argue that, if points in the subject and the intermediate images share similar local appearances, they may have common correspondence in the template. In this way, we learn the sparse representation of a certain subject point to reveal several similar candidate points in the intermediate images. Each selected intermediate candidate can bridge the correspondence from the subject point to the template space, thus predicting the transformation associated with the subject point at the confidence level that relates to the learned sparse coefficient. Following this strategy, we first predict transformations at selected key points, and retain multiple predictions on each key point, instead of allowing only a single correspondence. Then, by utilizing all key points and their predictions with varying confidences, we adaptively reconstruct the dense transformation field that warps the subject to the template. We further embed the prediction-reconstruction protocol above into a multi-resolution hierarchy. In the final, we refine our estimated transformation field via existing registration method in effective manners. We apply our method to registering brain MR images, and conclude that the proposed framework is competent to improve registration performances substantially.

AB - We propose a new approach to register the subject image with the template by leveraging a set of intermediate images that are pre-aligned to the template. We argue that, if points in the subject and the intermediate images share similar local appearances, they may have common correspondence in the template. In this way, we learn the sparse representation of a certain subject point to reveal several similar candidate points in the intermediate images. Each selected intermediate candidate can bridge the correspondence from the subject point to the template space, thus predicting the transformation associated with the subject point at the confidence level that relates to the learned sparse coefficient. Following this strategy, we first predict transformations at selected key points, and retain multiple predictions on each key point, instead of allowing only a single correspondence. Then, by utilizing all key points and their predictions with varying confidences, we adaptively reconstruct the dense transformation field that warps the subject to the template. We further embed the prediction-reconstruction protocol above into a multi-resolution hierarchy. In the final, we refine our estimated transformation field via existing registration method in effective manners. We apply our method to registering brain MR images, and conclude that the proposed framework is competent to improve registration performances substantially.

KW - Brain MR image registration

KW - Correspondence detection

KW - Deformable image registration

KW - Sparsity learning

KW - Transformation prediction

UR - http://www.scopus.com/inward/record.url?scp=84920889055&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84920889055&partnerID=8YFLogxK

U2 - 10.1016/j.media.2014.10.007

DO - 10.1016/j.media.2014.10.007

M3 - Article

C2 - 25476412

AN - SCOPUS:84920889055

VL - 20

SP - 61

EP - 75

JO - Medical Image Analysis

JF - Medical Image Analysis

SN - 1361-8415

IS - 1

ER -