Groupwise segmentation improves neuroimaging classification accuracy

Yaping Wang, Hongjun Jia, Pew Thian Yap, Bo Cheng, Chong Yaw Wee, Lei Guo, Dinggang Shen

Research output: Chapter in Book/Report/Conference proceedingConference contribution

4 Citations (Scopus)

Abstract

Accurate diagnosis of Alzheimer's disease (AD), especially mild cognitive impairment (MCI), is critical for treatment of the disease. Many algorithms have been proposed to improve classification performance. While most existing methods focus on exploring different feature extraction and selection techniques, in this paper, we show that the pre-processing steps for MRI scans, i.e., registration and segmentation, significantly affect the classification performance. Specifically, we evaluate the classification performance given by a multi-atlas based multi-image segmentation (MABMIS) method, with respect to more conventional segmentation methods. By incorporating tree-based groupwise registration and iterative groupwise segmentation strategies, MABMIS attains more accurate and consistent segmentation results compared with the conventional methods that do not take into account the inherent distribution of images under segmentation. This increased segmentation accuracy will benefit classification by minimizing errors that are propagated to the subsequent analysis steps. Experimental results indicate that MABMIS achieves better performance when compared with the conventional methods in the following classification tasks using the ADNI dataset: AD vs. MCI (accuracy: 71.8%), AD vs. healthy control (HC) (89.1%), progressive MCI vs. HC (84.4%), and progressive MCI vs. stable MCI (70.0%). These results show that pre-processing the images accurately is critical for neuroimaging classification.

Original languageEnglish
Title of host publicationLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Pages185-193
Number of pages9
Volume7509 LNCS
DOIs
Publication statusPublished - 2012 Nov 6
Externally publishedYes
Event2nd International Workshop on Multimodal Brain Image Analysis, MBIA 2012, Held in Conjunction with the 15th International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2012 - Nice, France
Duration: 2012 Oct 12012 Oct 5

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume7509 LNCS
ISSN (Print)03029743
ISSN (Electronic)16113349

Other

Other2nd International Workshop on Multimodal Brain Image Analysis, MBIA 2012, Held in Conjunction with the 15th International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2012
CountryFrance
CityNice
Period12/10/112/10/5

Fingerprint

Neuroimaging
Segmentation
Alzheimer's Disease
Atlas
Image segmentation
Image Segmentation
Registration
Preprocessing
Feature extraction
Processing
Feature Selection
Feature Extraction
Evaluate
Experimental Results

ASJC Scopus subject areas

  • Computer Science(all)
  • Theoretical Computer Science

Cite this

Wang, Y., Jia, H., Yap, P. T., Cheng, B., Wee, C. Y., Guo, L., & Shen, D. (2012). Groupwise segmentation improves neuroimaging classification accuracy. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 7509 LNCS, pp. 185-193). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 7509 LNCS). https://doi.org/10.1007/978-3-642-33530-3_16

Groupwise segmentation improves neuroimaging classification accuracy. / Wang, Yaping; Jia, Hongjun; Yap, Pew Thian; Cheng, Bo; Wee, Chong Yaw; Guo, Lei; Shen, Dinggang.

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 7509 LNCS 2012. p. 185-193 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 7509 LNCS).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Wang, Y, Jia, H, Yap, PT, Cheng, B, Wee, CY, Guo, L & Shen, D 2012, Groupwise segmentation improves neuroimaging classification accuracy. in Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). vol. 7509 LNCS, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 7509 LNCS, pp. 185-193, 2nd International Workshop on Multimodal Brain Image Analysis, MBIA 2012, Held in Conjunction with the 15th International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2012, Nice, France, 12/10/1. https://doi.org/10.1007/978-3-642-33530-3_16
Wang Y, Jia H, Yap PT, Cheng B, Wee CY, Guo L et al. Groupwise segmentation improves neuroimaging classification accuracy. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 7509 LNCS. 2012. p. 185-193. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-642-33530-3_16
Wang, Yaping ; Jia, Hongjun ; Yap, Pew Thian ; Cheng, Bo ; Wee, Chong Yaw ; Guo, Lei ; Shen, Dinggang. / Groupwise segmentation improves neuroimaging classification accuracy. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 7509 LNCS 2012. pp. 185-193 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{6d21bd80bbb04945be0b7438e626ed47,
title = "Groupwise segmentation improves neuroimaging classification accuracy",
abstract = "Accurate diagnosis of Alzheimer's disease (AD), especially mild cognitive impairment (MCI), is critical for treatment of the disease. Many algorithms have been proposed to improve classification performance. While most existing methods focus on exploring different feature extraction and selection techniques, in this paper, we show that the pre-processing steps for MRI scans, i.e., registration and segmentation, significantly affect the classification performance. Specifically, we evaluate the classification performance given by a multi-atlas based multi-image segmentation (MABMIS) method, with respect to more conventional segmentation methods. By incorporating tree-based groupwise registration and iterative groupwise segmentation strategies, MABMIS attains more accurate and consistent segmentation results compared with the conventional methods that do not take into account the inherent distribution of images under segmentation. This increased segmentation accuracy will benefit classification by minimizing errors that are propagated to the subsequent analysis steps. Experimental results indicate that MABMIS achieves better performance when compared with the conventional methods in the following classification tasks using the ADNI dataset: AD vs. MCI (accuracy: 71.8{\%}), AD vs. healthy control (HC) (89.1{\%}), progressive MCI vs. HC (84.4{\%}), and progressive MCI vs. stable MCI (70.0{\%}). These results show that pre-processing the images accurately is critical for neuroimaging classification.",
author = "Yaping Wang and Hongjun Jia and Yap, {Pew Thian} and Bo Cheng and Wee, {Chong Yaw} and Lei Guo and Dinggang Shen",
year = "2012",
month = "11",
day = "6",
doi = "10.1007/978-3-642-33530-3_16",
language = "English",
isbn = "9783642335297",
volume = "7509 LNCS",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
pages = "185--193",
booktitle = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

}

TY - GEN

T1 - Groupwise segmentation improves neuroimaging classification accuracy

AU - Wang, Yaping

AU - Jia, Hongjun

AU - Yap, Pew Thian

AU - Cheng, Bo

AU - Wee, Chong Yaw

AU - Guo, Lei

AU - Shen, Dinggang

PY - 2012/11/6

Y1 - 2012/11/6

N2 - Accurate diagnosis of Alzheimer's disease (AD), especially mild cognitive impairment (MCI), is critical for treatment of the disease. Many algorithms have been proposed to improve classification performance. While most existing methods focus on exploring different feature extraction and selection techniques, in this paper, we show that the pre-processing steps for MRI scans, i.e., registration and segmentation, significantly affect the classification performance. Specifically, we evaluate the classification performance given by a multi-atlas based multi-image segmentation (MABMIS) method, with respect to more conventional segmentation methods. By incorporating tree-based groupwise registration and iterative groupwise segmentation strategies, MABMIS attains more accurate and consistent segmentation results compared with the conventional methods that do not take into account the inherent distribution of images under segmentation. This increased segmentation accuracy will benefit classification by minimizing errors that are propagated to the subsequent analysis steps. Experimental results indicate that MABMIS achieves better performance when compared with the conventional methods in the following classification tasks using the ADNI dataset: AD vs. MCI (accuracy: 71.8%), AD vs. healthy control (HC) (89.1%), progressive MCI vs. HC (84.4%), and progressive MCI vs. stable MCI (70.0%). These results show that pre-processing the images accurately is critical for neuroimaging classification.

AB - Accurate diagnosis of Alzheimer's disease (AD), especially mild cognitive impairment (MCI), is critical for treatment of the disease. Many algorithms have been proposed to improve classification performance. While most existing methods focus on exploring different feature extraction and selection techniques, in this paper, we show that the pre-processing steps for MRI scans, i.e., registration and segmentation, significantly affect the classification performance. Specifically, we evaluate the classification performance given by a multi-atlas based multi-image segmentation (MABMIS) method, with respect to more conventional segmentation methods. By incorporating tree-based groupwise registration and iterative groupwise segmentation strategies, MABMIS attains more accurate and consistent segmentation results compared with the conventional methods that do not take into account the inherent distribution of images under segmentation. This increased segmentation accuracy will benefit classification by minimizing errors that are propagated to the subsequent analysis steps. Experimental results indicate that MABMIS achieves better performance when compared with the conventional methods in the following classification tasks using the ADNI dataset: AD vs. MCI (accuracy: 71.8%), AD vs. healthy control (HC) (89.1%), progressive MCI vs. HC (84.4%), and progressive MCI vs. stable MCI (70.0%). These results show that pre-processing the images accurately is critical for neuroimaging classification.

UR - http://www.scopus.com/inward/record.url?scp=84868237036&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84868237036&partnerID=8YFLogxK

U2 - 10.1007/978-3-642-33530-3_16

DO - 10.1007/978-3-642-33530-3_16

M3 - Conference contribution

AN - SCOPUS:84868237036

SN - 9783642335297

VL - 7509 LNCS

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 185

EP - 193

BT - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

ER -