Cerebellum segmentation in MRI using atlas registration and local multi-scale image descriptors

F. van der Lijn; M. de Bruijne; Y.Y. Hoogendam; S. Klein; R. Hameeteman; M. Breteler; W. Niessen

doi:10.1109/ISBI.2009.5193023

Cerebellum segmentation in MRI using atlas registration and local multi-scale image descriptors

F. van der Lijn, M. de Bruijne, Y.Y. Hoogendam, S. Klein, R. Hameeteman, M. Breteler, W. Niessen

Department of Computer Science

Research output: Chapter in Book/Report/Conference proceeding › Article in proceedings › Research › peer-review

21 Citations (Scopus)

Abstract

We propose a novel cerebellum segmentation method for MRI, based on a combination of statistical models of the structure's expected location in the brain and its local appearance. The appearance model is obtained from a k-nearest-neighbor classifier, which uses a set of multi-scale local image descriptors as features. The spatial model is constructed by registering multiple manually annotated datasets to the unlabeled target image. The two components are then combined in a Bayesian framework. The method is quantitatively validated in a leave-one-out experiment using 18 MR images of elderly subjects. The experiment showed that the method produces accurate segmentations. The mean Dice similarity index compared to the manual reference was 0.953 for left and right, and the mean surface distance was 0.49 mm for left and 0.50 mm for right. The combined atlas- and appearance-based method was found to be more accurate than a method based on atlas-registration alone.

Original language	English
Title of host publication	IEEE International Symposium on Biomedical Imaging (ISBI'09) : From Nano to Macro
Number of pages	4
Publication date	2009
Pages	221-224
ISBN (Print)	978-1-4244-3931-7
DOIs	https://doi.org/10.1109/ISBI.2009.5193023
Publication status	Published - 2009
Event	ISBI 2009, 6th IEEE International Symposium on Biomedical Imaging: From Nano to Macro - Boston, Massachusetts, United States Duration: 28 Jun 0009 → 1 Jul 0009 Conference number: 6

Conference

Conference	ISBI 2009, 6th IEEE International Symposium on Biomedical Imaging: From Nano to Macro
Number	6
Country/Territory	United States
City	Boston, Massachusetts
Period	28/06/0009 → 01/07/0009

Series	Uden navn
ISSN	1945-7928

Access to Document

10.1109/ISBI.2009.5193023

Cite this

Cerebellum segmentation in MRI using atlas registration and local multi-scale image descriptors. / van der Lijn, F.; de Bruijne, M.; Hoogendam, Y.Y.; Klein, S.; Hameeteman, R.; Breteler, M.; Niessen, W.

IEEE International Symposium on Biomedical Imaging (ISBI'09): From Nano to Macro. 2009. p. 221-224 (Uden navn).

Research output: Chapter in Book/Report/Conference proceeding › Article in proceedings › Research › peer-review

van der Lijn, F, de Bruijne, M, Hoogendam, YY, Klein, S, Hameeteman, R, Breteler, M & Niessen, W 2009, Cerebellum segmentation in MRI using atlas registration and local multi-scale image descriptors. in IEEE International Symposium on Biomedical Imaging (ISBI'09): From Nano to Macro. Uden navn, pp. 221-224, ISBI 2009, 6th IEEE International Symposium on Biomedical Imaging: From Nano to Macro, Boston, Massachusetts, United States, 28/06/0009. https://doi.org/10.1109/ISBI.2009.5193023

@inproceedings{b0b86bb09d3411debc73000ea68e967b,

title = "Cerebellum segmentation in MRI using atlas registration and local multi-scale image descriptors",

abstract = "We propose a novel cerebellum segmentation method for MRI, based on a combination of statistical models of the structure's expected location in the brain and its local appearance. The appearance model is obtained from a k-nearest-neighbor classifier, which uses a set of multi-scale local image descriptors as features. The spatial model is constructed by registering multiple manually annotated datasets to the unlabeled target image. The two components are then combined in a Bayesian framework. The method is quantitatively validated in a leave-one-out experiment using 18 MR images of elderly subjects. The experiment showed that the method produces accurate segmentations. The mean Dice similarity index compared to the manual reference was 0.953 for left and right, and the mean surface distance was 0.49 mm for left and 0.50 mm for right. The combined atlas- and appearance-based method was found to be more accurate than a method based on atlas-registration alone.",

author = "{van der Lijn}, F. and {de Bruijne}, M. and Y.Y. Hoogendam and S. Klein and R. Hameeteman and M. Breteler and W. Niessen",

year = "2009",

doi = "10.1109/ISBI.2009.5193023",

language = "English",

isbn = "978-1-4244-3931-7",

series = "Uden navn",

pages = "221--224",

booktitle = "IEEE International Symposium on Biomedical Imaging (ISBI'09)",

note = "null ; Conference date: 28-06-0009 Through 01-07-0009",

}

TY - GEN

T1 - Cerebellum segmentation in MRI using atlas registration and local multi-scale image descriptors

AU - van der Lijn, F.

AU - de Bruijne, M.

AU - Hoogendam, Y.Y.

AU - Klein, S.

AU - Hameeteman, R.

AU - Breteler, M.

AU - Niessen, W.

N1 - Conference code: 6

PY - 2009

Y1 - 2009

N2 - We propose a novel cerebellum segmentation method for MRI, based on a combination of statistical models of the structure's expected location in the brain and its local appearance. The appearance model is obtained from a k-nearest-neighbor classifier, which uses a set of multi-scale local image descriptors as features. The spatial model is constructed by registering multiple manually annotated datasets to the unlabeled target image. The two components are then combined in a Bayesian framework. The method is quantitatively validated in a leave-one-out experiment using 18 MR images of elderly subjects. The experiment showed that the method produces accurate segmentations. The mean Dice similarity index compared to the manual reference was 0.953 for left and right, and the mean surface distance was 0.49 mm for left and 0.50 mm for right. The combined atlas- and appearance-based method was found to be more accurate than a method based on atlas-registration alone.

AB - We propose a novel cerebellum segmentation method for MRI, based on a combination of statistical models of the structure's expected location in the brain and its local appearance. The appearance model is obtained from a k-nearest-neighbor classifier, which uses a set of multi-scale local image descriptors as features. The spatial model is constructed by registering multiple manually annotated datasets to the unlabeled target image. The two components are then combined in a Bayesian framework. The method is quantitatively validated in a leave-one-out experiment using 18 MR images of elderly subjects. The experiment showed that the method produces accurate segmentations. The mean Dice similarity index compared to the manual reference was 0.953 for left and right, and the mean surface distance was 0.49 mm for left and 0.50 mm for right. The combined atlas- and appearance-based method was found to be more accurate than a method based on atlas-registration alone.

U2 - 10.1109/ISBI.2009.5193023

DO - 10.1109/ISBI.2009.5193023

M3 - Article in proceedings

SN - 978-1-4244-3931-7

T3 - Uden navn

SP - 221

EP - 224

BT - IEEE International Symposium on Biomedical Imaging (ISBI'09)

Y2 - 28 June 0009 through 1 July 0009

ER -