Fast and robust multi-atlas segmentation of brain magnetic resonance images

TY - JOUR

T1 - Fast and robust multi-atlas segmentation of brain magnetic resonance images

AU - Lötjönen, Jyrki Mp

AU - Wolz, Robin

AU - Koikkalainen, Juha R

AU - Thurfjell, Lennart

AU - Waldemar, Gunhild

AU - Soininen, Hilkka

AU - Rueckert, Daniel

AU - Alzheimer's Disease Neuroimaging Initiative

N1 - Copyright (c) 2009 Elsevier Inc. All rights reserved.

PY - 2010/2/1

Y1 - 2010/2/1

N2 - We introduce an optimised pipeline for multi-atlas brain MRI segmentation. Both accuracy and speed of segmentation are considered. We study different similarity measures used in non-rigid registration. We show that intensity differences for intensity normalised images can be used instead of standard normalised mutual information in registration without compromising the accuracy but leading to threefold decrease in the computation time. We study and validate also different methods for atlas selection. Finally, we propose two new approaches for combining multi-atlas segmentation and intensity modelling based on segmentation using expectation maximisation (EM) and optimisation via graph cuts. The segmentation pipeline is evaluated with two data cohorts: IBSR data (N=18, six subcortial structures: thalamus, caudate, putamen, pallidum, hippocampus, amygdala) and ADNI data (N=60, hippocampus). The average similarity index between automatically and manually generated volumes was 0.849 (IBSR, six subcortical structures) and 0.880 (ADNI, hippocampus). The correlation coefficient for hippocampal volumes was 0.95 with the ADNI data. The computation time using a standard multicore PC computer was about 3-4 min. Our results compare favourably with other recently published results.

AB - We introduce an optimised pipeline for multi-atlas brain MRI segmentation. Both accuracy and speed of segmentation are considered. We study different similarity measures used in non-rigid registration. We show that intensity differences for intensity normalised images can be used instead of standard normalised mutual information in registration without compromising the accuracy but leading to threefold decrease in the computation time. We study and validate also different methods for atlas selection. Finally, we propose two new approaches for combining multi-atlas segmentation and intensity modelling based on segmentation using expectation maximisation (EM) and optimisation via graph cuts. The segmentation pipeline is evaluated with two data cohorts: IBSR data (N=18, six subcortial structures: thalamus, caudate, putamen, pallidum, hippocampus, amygdala) and ADNI data (N=60, hippocampus). The average similarity index between automatically and manually generated volumes was 0.849 (IBSR, six subcortical structures) and 0.880 (ADNI, hippocampus). The correlation coefficient for hippocampal volumes was 0.95 with the ADNI data. The computation time using a standard multicore PC computer was about 3-4 min. Our results compare favourably with other recently published results.

U2 - 10.1016/j.neuroimage.2009.10.026

DO - 10.1016/j.neuroimage.2009.10.026

M3 - Journal article

SN - 1053-8119

VL - 49

SP - 2352

EP - 2365

JO - NeuroImage

JF - NeuroImage

IS - 3

ER -