Inter-modality relationship constrained multi-task feature selection for AD/MCI classification

Feng Liu, Chong Yaw Wee, Huafu Chen, Dinggang Shen

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

Abstract

In conventional multi-modality based classification framework, feature selection is typically performed separately for each individual modality, ignoring potential strong inter-modality relationship of the same subject. To extract this inter-modality relationship, L2,1 norm-based multi-task learning approach can be used to jointly select common features from different modalities. Unfortunately, this approach overlooks different yet complementary information conveyed by different modalities. To address this issue, we propose a novel multi-task feature selection method to effectively preserve the complementary information between different modalities, improving brain disease classification accuracy. Specifically, a new constraint is introduced to preserve the inter-modality relationship by treating the feature selection procedure of each modality as a task. This constraint preserves distance between feature vectors from different modalities after projection to low dimensional feature space. We evaluated our method on the Alzheimer's Disease Neuroimaging Initiative (ADNI) dataset and obtained significant improvement on Alzheimer's Disease (AD) and Mild Cognitive Impairment (MCI) classification compared to state-of-the-art methods.

Original languageEnglish
Title of host publicationLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Pages308-315
Number of pages8
Volume8149 LNCS
EditionPART 1
DOIs
Publication statusPublished - 2013 Oct 23
Externally publishedYes
Event16th International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2013 - Nagoya, Japan
Duration: 2013 Sep 222013 Sep 26

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
NumberPART 1
Volume8149 LNCS
ISSN (Print)03029743
ISSN (Electronic)16113349

Other

Other16th International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2013
CountryJapan
CityNagoya
Period13/9/2213/9/26

Fingerprint

Alzheimer's Disease
Feature Selection
Modality
Feature extraction
Neuroimaging
Brain
Multi-task Learning
Relationships
Multimodality
Selection Procedures
Feature Space
Feature Vector
Projection
Norm

Keywords

  • Alzheimer's Disease
  • Multi-kernel support vector machine
  • Multi-modality
  • Multi-task learning
  • Sparse representation

ASJC Scopus subject areas

  • Computer Science(all)
  • Theoretical Computer Science

Cite this

Liu, F., Wee, C. Y., Chen, H., & Shen, D. (2013). Inter-modality relationship constrained multi-task feature selection for AD/MCI classification. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (PART 1 ed., Vol. 8149 LNCS, pp. 308-315). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 8149 LNCS, No. PART 1). https://doi.org/10.1007/978-3-642-40811-3_39

Inter-modality relationship constrained multi-task feature selection for AD/MCI classification. / Liu, Feng; Wee, Chong Yaw; Chen, Huafu; Shen, Dinggang.

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 8149 LNCS PART 1. ed. 2013. p. 308-315 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 8149 LNCS, No. PART 1).

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

Liu, F, Wee, CY, Chen, H & Shen, D 2013, Inter-modality relationship constrained multi-task feature selection for AD/MCI classification. in Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). PART 1 edn, vol. 8149 LNCS, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), no. PART 1, vol. 8149 LNCS, pp. 308-315, 16th International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2013, Nagoya, Japan, 13/9/22. https://doi.org/10.1007/978-3-642-40811-3_39
Liu F, Wee CY, Chen H, Shen D. Inter-modality relationship constrained multi-task feature selection for AD/MCI classification. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). PART 1 ed. Vol. 8149 LNCS. 2013. p. 308-315. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); PART 1). https://doi.org/10.1007/978-3-642-40811-3_39
Liu, Feng ; Wee, Chong Yaw ; Chen, Huafu ; Shen, Dinggang. / Inter-modality relationship constrained multi-task feature selection for AD/MCI classification. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 8149 LNCS PART 1. ed. 2013. pp. 308-315 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); PART 1).
@inproceedings{65ab2e352bb3437d9941f00d293f9fd5,
title = "Inter-modality relationship constrained multi-task feature selection for AD/MCI classification",
abstract = "In conventional multi-modality based classification framework, feature selection is typically performed separately for each individual modality, ignoring potential strong inter-modality relationship of the same subject. To extract this inter-modality relationship, L2,1 norm-based multi-task learning approach can be used to jointly select common features from different modalities. Unfortunately, this approach overlooks different yet complementary information conveyed by different modalities. To address this issue, we propose a novel multi-task feature selection method to effectively preserve the complementary information between different modalities, improving brain disease classification accuracy. Specifically, a new constraint is introduced to preserve the inter-modality relationship by treating the feature selection procedure of each modality as a task. This constraint preserves distance between feature vectors from different modalities after projection to low dimensional feature space. We evaluated our method on the Alzheimer's Disease Neuroimaging Initiative (ADNI) dataset and obtained significant improvement on Alzheimer's Disease (AD) and Mild Cognitive Impairment (MCI) classification compared to state-of-the-art methods.",
keywords = "Alzheimer's Disease, Multi-kernel support vector machine, Multi-modality, Multi-task learning, Sparse representation",
author = "Feng Liu and Wee, {Chong Yaw} and Huafu Chen and Dinggang Shen",
year = "2013",
month = "10",
day = "23",
doi = "10.1007/978-3-642-40811-3_39",
language = "English",
isbn = "9783642408106",
volume = "8149 LNCS",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
number = "PART 1",
pages = "308--315",
booktitle = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
edition = "PART 1",

}

TY - GEN

T1 - Inter-modality relationship constrained multi-task feature selection for AD/MCI classification

AU - Liu, Feng

AU - Wee, Chong Yaw

AU - Chen, Huafu

AU - Shen, Dinggang

PY - 2013/10/23

Y1 - 2013/10/23

N2 - In conventional multi-modality based classification framework, feature selection is typically performed separately for each individual modality, ignoring potential strong inter-modality relationship of the same subject. To extract this inter-modality relationship, L2,1 norm-based multi-task learning approach can be used to jointly select common features from different modalities. Unfortunately, this approach overlooks different yet complementary information conveyed by different modalities. To address this issue, we propose a novel multi-task feature selection method to effectively preserve the complementary information between different modalities, improving brain disease classification accuracy. Specifically, a new constraint is introduced to preserve the inter-modality relationship by treating the feature selection procedure of each modality as a task. This constraint preserves distance between feature vectors from different modalities after projection to low dimensional feature space. We evaluated our method on the Alzheimer's Disease Neuroimaging Initiative (ADNI) dataset and obtained significant improvement on Alzheimer's Disease (AD) and Mild Cognitive Impairment (MCI) classification compared to state-of-the-art methods.

AB - In conventional multi-modality based classification framework, feature selection is typically performed separately for each individual modality, ignoring potential strong inter-modality relationship of the same subject. To extract this inter-modality relationship, L2,1 norm-based multi-task learning approach can be used to jointly select common features from different modalities. Unfortunately, this approach overlooks different yet complementary information conveyed by different modalities. To address this issue, we propose a novel multi-task feature selection method to effectively preserve the complementary information between different modalities, improving brain disease classification accuracy. Specifically, a new constraint is introduced to preserve the inter-modality relationship by treating the feature selection procedure of each modality as a task. This constraint preserves distance between feature vectors from different modalities after projection to low dimensional feature space. We evaluated our method on the Alzheimer's Disease Neuroimaging Initiative (ADNI) dataset and obtained significant improvement on Alzheimer's Disease (AD) and Mild Cognitive Impairment (MCI) classification compared to state-of-the-art methods.

KW - Alzheimer's Disease

KW - Multi-kernel support vector machine

KW - Multi-modality

KW - Multi-task learning

KW - Sparse representation

UR - http://www.scopus.com/inward/record.url?scp=84885776074&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84885776074&partnerID=8YFLogxK

U2 - 10.1007/978-3-642-40811-3_39

DO - 10.1007/978-3-642-40811-3_39

M3 - Conference contribution

SN - 9783642408106

VL - 8149 LNCS

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

SP - 308

EP - 315

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

ER -