Automated segmentation of the corpus callosum in midsagittal brain magnetic resonance images

Chul Hee Lee, Shin Huh, Terence A. Ketter, Michael Unser

Research output: Contribution to journalArticle

10 Citations (Scopus)

Abstract

We propose a new algorithm to find the corpus callosum automatically from midsagittal brain MR (magnetic resonance) images using the statistical characteristics and shape information of the corpus callosum. We first extract regions satisfying the statistical characteristics (gray level distributions) of the corpus callosum that have relatively high intensity values. Then we try to find a region matching the shape information of the corpus callosum. In order to match the shape information, we propose a new directed window region growing algorithm instead of using conventional contour matching. An innovative feature of the algorithm is that we adaptively relax the statistical requirement until we find a region matching the shape information. After the initial segmentation, a directed border path pruning algorithm is proposed in order to remove some undesired artifacts, especially on the top of the corpus callosum. The proposed algorithm was applied to over 120 images and provided promising results.

Original languageEnglish
Pages (from-to)924-935
Number of pages12
JournalOptical Engineering
Volume39
Issue number4
DOIs
Publication statusPublished - 2000 Apr 1

Fingerprint

Magnetic resonance
brain
magnetic resonance
Brain
borders
artifacts
requirements

All Science Journal Classification (ASJC) codes

  • Atomic and Molecular Physics, and Optics

Cite this

Lee, Chul Hee ; Huh, Shin ; Ketter, Terence A. ; Unser, Michael. / Automated segmentation of the corpus callosum in midsagittal brain magnetic resonance images. In: Optical Engineering. 2000 ; Vol. 39, No. 4. pp. 924-935.
@article{da04f95909f94078ad7e90d5e197920b,
title = "Automated segmentation of the corpus callosum in midsagittal brain magnetic resonance images",
abstract = "We propose a new algorithm to find the corpus callosum automatically from midsagittal brain MR (magnetic resonance) images using the statistical characteristics and shape information of the corpus callosum. We first extract regions satisfying the statistical characteristics (gray level distributions) of the corpus callosum that have relatively high intensity values. Then we try to find a region matching the shape information of the corpus callosum. In order to match the shape information, we propose a new directed window region growing algorithm instead of using conventional contour matching. An innovative feature of the algorithm is that we adaptively relax the statistical requirement until we find a region matching the shape information. After the initial segmentation, a directed border path pruning algorithm is proposed in order to remove some undesired artifacts, especially on the top of the corpus callosum. The proposed algorithm was applied to over 120 images and provided promising results.",
author = "Lee, {Chul Hee} and Shin Huh and Ketter, {Terence A.} and Michael Unser",
year = "2000",
month = "4",
day = "1",
doi = "10.1117/1.602449",
language = "English",
volume = "39",
pages = "924--935",
journal = "Optical Engineering",
issn = "0091-3286",
publisher = "SPIE",
number = "4",

}

Automated segmentation of the corpus callosum in midsagittal brain magnetic resonance images. / Lee, Chul Hee; Huh, Shin; Ketter, Terence A.; Unser, Michael.

In: Optical Engineering, Vol. 39, No. 4, 01.04.2000, p. 924-935.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Automated segmentation of the corpus callosum in midsagittal brain magnetic resonance images

AU - Lee, Chul Hee

AU - Huh, Shin

AU - Ketter, Terence A.

AU - Unser, Michael

PY - 2000/4/1

Y1 - 2000/4/1

N2 - We propose a new algorithm to find the corpus callosum automatically from midsagittal brain MR (magnetic resonance) images using the statistical characteristics and shape information of the corpus callosum. We first extract regions satisfying the statistical characteristics (gray level distributions) of the corpus callosum that have relatively high intensity values. Then we try to find a region matching the shape information of the corpus callosum. In order to match the shape information, we propose a new directed window region growing algorithm instead of using conventional contour matching. An innovative feature of the algorithm is that we adaptively relax the statistical requirement until we find a region matching the shape information. After the initial segmentation, a directed border path pruning algorithm is proposed in order to remove some undesired artifacts, especially on the top of the corpus callosum. The proposed algorithm was applied to over 120 images and provided promising results.

AB - We propose a new algorithm to find the corpus callosum automatically from midsagittal brain MR (magnetic resonance) images using the statistical characteristics and shape information of the corpus callosum. We first extract regions satisfying the statistical characteristics (gray level distributions) of the corpus callosum that have relatively high intensity values. Then we try to find a region matching the shape information of the corpus callosum. In order to match the shape information, we propose a new directed window region growing algorithm instead of using conventional contour matching. An innovative feature of the algorithm is that we adaptively relax the statistical requirement until we find a region matching the shape information. After the initial segmentation, a directed border path pruning algorithm is proposed in order to remove some undesired artifacts, especially on the top of the corpus callosum. The proposed algorithm was applied to over 120 images and provided promising results.

UR - http://www.scopus.com/inward/record.url?scp=0033743351&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0033743351&partnerID=8YFLogxK

U2 - 10.1117/1.602449

DO - 10.1117/1.602449

M3 - Article

VL - 39

SP - 924

EP - 935

JO - Optical Engineering

JF - Optical Engineering

SN - 0091-3286

IS - 4

ER -