Please contact IT if you would like this publication removed.

Journal Article

Title

A robust method for extraction and automatic segmentation of brain images

Journal Name

NeuroImage

Abstract

A new protocol is introduced for brain extraction and automatic tissuesegmentation of MR images. For the brain extraction algorithm, proton densityand T2-weighted images are used to generate a brain mask encompassing the fullintracranial cavity. Segmentation of brain tissues into gray matter (GM), whitematter (WM), and cerebral spinal fluid (CSF) is accomplished on a T1-weightedimage after applying the brain mask. The fully automatic segmentation algorithmis histogram-based and uses the Expectation Maximization algorithm to model afour-Gaussian mixture for both global and local histograms. The means of thelocal Gaussians for GM, WM, and CSF are used to set local thresholds for tissueclassification. Reproducibility of the extraction procedure was excellent, withaverage variation in intracranial capacity (TIC) of 0.13 and 0.66% TIC in 12healthy normal and 33 Alzheimer brains, respectively. Repeatability of thesegmentation algorithm, tested on healthy normal images, indicated scan-rescandifferences in global tissue volumes of less than 0.30% TIC. Reproducibility atthe regional level was established by comparing segmentation results within the12 major Talairach subdivisions. Accuracy of the algorithm was tested on adigital brain phantom, and errors were less than 1% of the phantom volume.Maximal Type I and Type II classification errors were low, ranging between 2.2and 4.3% of phantom volume. The algorithm was also insensitive to variation inparameter initialization values. The protocol is robust, fast, and its successin segmenting normal as well as diseased brains makes it an attractive clinicalapplication.

Volume

17

Year

2002

Pages

1087-1100

Authors

Kovacevic N., Lobaugh N.J., Bronskill M.J., Levine B., Feinstein A. & Black S.E.

Download the Article