Robust anatomical landmark detection with application to MR brain image registration

Dong Han, Yaozong Gao, Guorong Wu, Pew Thian Yap, Dinggang Shen

Research output: Contribution to journalArticle

13 Citations (Scopus)

Abstract

Comparison of human brain MR images is often challenged by large inter-subject structural variability. To determine correspondences between MR brain images, most existing methods typically perform a local neighborhood search, based on certain morphological features. They are limited in two aspects: (1) pre-defined morphological features often have limited power in characterizing brain structures, thus leading to inaccurate correspondence detection, and (2) correspondence matching is often restricted within local small neighborhoods and fails to cater to images with large anatomical difference. To address these limitations, we propose a novel method to detect distinctive landmarks for effective correspondence matching. Specifically, we first annotate a group of landmarks in a large set of training MR brain images. Then, we use regression forest to simultaneously learn (1) the optimal sets of features to best characterize each landmark and (2) the non-linear mappings from the local patch appearances of image points to their 3D displacements towards each landmark. The learned regression forests are used as landmark detectors to predict the locations of these landmarks in new images. Because each detector is learned based on features that best distinguish the landmark from other points and also landmark detection is performed in the entire image domain, our method can address the limitations in conventional methods. The deformation field estimated based on the alignment of these detected landmarks can then be used as initialization for image registration. Experimental results show that our method is capable of providing good initialization even for the images with large deformation difference, thus improving registration accuracy.

Original languageEnglish
Pages (from-to)277-290
Number of pages14
JournalComputerized Medical Imaging and Graphics
Volume46
DOIs
Publication statusPublished - 2015 Dec 1

Fingerprint

Image registration
Brain
Detectors
Forests

Keywords

  • Anatomical landmark detection
  • Brain MRI
  • Deformable registration
  • Random forest regression

ASJC Scopus subject areas

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

Cite this

Robust anatomical landmark detection with application to MR brain image registration. / Han, Dong; Gao, Yaozong; Wu, Guorong; Yap, Pew Thian; Shen, Dinggang.

In: Computerized Medical Imaging and Graphics, Vol. 46, 01.12.2015, p. 277-290.

Research output: Contribution to journalArticle

Han, Dong ; Gao, Yaozong ; Wu, Guorong ; Yap, Pew Thian ; Shen, Dinggang. / Robust anatomical landmark detection with application to MR brain image registration. In: Computerized Medical Imaging and Graphics. 2015 ; Vol. 46. pp. 277-290.
@article{a5bc6e5a1fa7469b96f62968b6f3ac7b,
title = "Robust anatomical landmark detection with application to MR brain image registration",
abstract = "Comparison of human brain MR images is often challenged by large inter-subject structural variability. To determine correspondences between MR brain images, most existing methods typically perform a local neighborhood search, based on certain morphological features. They are limited in two aspects: (1) pre-defined morphological features often have limited power in characterizing brain structures, thus leading to inaccurate correspondence detection, and (2) correspondence matching is often restricted within local small neighborhoods and fails to cater to images with large anatomical difference. To address these limitations, we propose a novel method to detect distinctive landmarks for effective correspondence matching. Specifically, we first annotate a group of landmarks in a large set of training MR brain images. Then, we use regression forest to simultaneously learn (1) the optimal sets of features to best characterize each landmark and (2) the non-linear mappings from the local patch appearances of image points to their 3D displacements towards each landmark. The learned regression forests are used as landmark detectors to predict the locations of these landmarks in new images. Because each detector is learned based on features that best distinguish the landmark from other points and also landmark detection is performed in the entire image domain, our method can address the limitations in conventional methods. The deformation field estimated based on the alignment of these detected landmarks can then be used as initialization for image registration. Experimental results show that our method is capable of providing good initialization even for the images with large deformation difference, thus improving registration accuracy.",
keywords = "Anatomical landmark detection, Brain MRI, Deformable registration, Random forest regression",
author = "Dong Han and Yaozong Gao and Guorong Wu and Yap, {Pew Thian} and Dinggang Shen",
year = "2015",
month = "12",
day = "1",
doi = "10.1016/j.compmedimag.2015.09.002",
language = "English",
volume = "46",
pages = "277--290",
journal = "Computerized Medical Imaging and Graphics",
issn = "0895-6111",
publisher = "Elsevier Limited",

}

TY - JOUR

T1 - Robust anatomical landmark detection with application to MR brain image registration

AU - Han, Dong

AU - Gao, Yaozong

AU - Wu, Guorong

AU - Yap, Pew Thian

AU - Shen, Dinggang

PY - 2015/12/1

Y1 - 2015/12/1

N2 - Comparison of human brain MR images is often challenged by large inter-subject structural variability. To determine correspondences between MR brain images, most existing methods typically perform a local neighborhood search, based on certain morphological features. They are limited in two aspects: (1) pre-defined morphological features often have limited power in characterizing brain structures, thus leading to inaccurate correspondence detection, and (2) correspondence matching is often restricted within local small neighborhoods and fails to cater to images with large anatomical difference. To address these limitations, we propose a novel method to detect distinctive landmarks for effective correspondence matching. Specifically, we first annotate a group of landmarks in a large set of training MR brain images. Then, we use regression forest to simultaneously learn (1) the optimal sets of features to best characterize each landmark and (2) the non-linear mappings from the local patch appearances of image points to their 3D displacements towards each landmark. The learned regression forests are used as landmark detectors to predict the locations of these landmarks in new images. Because each detector is learned based on features that best distinguish the landmark from other points and also landmark detection is performed in the entire image domain, our method can address the limitations in conventional methods. The deformation field estimated based on the alignment of these detected landmarks can then be used as initialization for image registration. Experimental results show that our method is capable of providing good initialization even for the images with large deformation difference, thus improving registration accuracy.

AB - Comparison of human brain MR images is often challenged by large inter-subject structural variability. To determine correspondences between MR brain images, most existing methods typically perform a local neighborhood search, based on certain morphological features. They are limited in two aspects: (1) pre-defined morphological features often have limited power in characterizing brain structures, thus leading to inaccurate correspondence detection, and (2) correspondence matching is often restricted within local small neighborhoods and fails to cater to images with large anatomical difference. To address these limitations, we propose a novel method to detect distinctive landmarks for effective correspondence matching. Specifically, we first annotate a group of landmarks in a large set of training MR brain images. Then, we use regression forest to simultaneously learn (1) the optimal sets of features to best characterize each landmark and (2) the non-linear mappings from the local patch appearances of image points to their 3D displacements towards each landmark. The learned regression forests are used as landmark detectors to predict the locations of these landmarks in new images. Because each detector is learned based on features that best distinguish the landmark from other points and also landmark detection is performed in the entire image domain, our method can address the limitations in conventional methods. The deformation field estimated based on the alignment of these detected landmarks can then be used as initialization for image registration. Experimental results show that our method is capable of providing good initialization even for the images with large deformation difference, thus improving registration accuracy.

KW - Anatomical landmark detection

KW - Brain MRI

KW - Deformable registration

KW - Random forest regression

UR - http://www.scopus.com/inward/record.url?scp=84983209094&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84983209094&partnerID=8YFLogxK

U2 - 10.1016/j.compmedimag.2015.09.002

DO - 10.1016/j.compmedimag.2015.09.002

M3 - Article

C2 - 26433614

AN - SCOPUS:84983209094

VL - 46

SP - 277

EP - 290

JO - Computerized Medical Imaging and Graphics

JF - Computerized Medical Imaging and Graphics

SN - 0895-6111

ER -