Deep Learning-Based Regional Sub-models Integration for Parkinson’s Disease Diagnosis Using Diffusion Tensor Imaging

Hengling Zhao; Chih Chien Tsai; Ce Zhu; Mingyi Zhou; Jiun Jie Wang; Yipeng Liu

doi:10.1007/978-3-030-87358-5_60

Deep Learning-Based Regional Sub-models Integration for Parkinson’s Disease Diagnosis Using Diffusion Tensor Imaging

Hengling Zhao, Chih Chien Tsai, Ce Zhu, Mingyi Zhou, Jiun Jie Wang, Yipeng Liu^*

^*Corresponding author for this work

Department of Medical Imaging and Radiological Sciences

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

Abstract

Parkinson’s disease (PD) is a neurodegenerative disease. PD patients may have serious movement disorders and mental problems. The current diagnosis requires a professionally trained medical doctor to take a long period for it. Different doctors may even have different accuracies. Recently advances in deep learning-based medical image classification make it is possible to diagnose PD automatically. Different from most of the existing works on magnetic resonance images, we use diffusion tensor imaging (DTI) in that it can reflect functional data of the brain. We propose a sub-models integration framework based on convolutional neural networks (CNNs) for Parkinson’s disease. Each sub-region of the brain is used to train a unique CNN model, named sub-model, and the selective stacking algorithm is used to screen these sub-models. It obtains the classification accuracy of 92.4% on the cross-validation dataset. In addition, it can provide that which sub-regions play a role in the judgment of the final result so that this framework has stronger practical application than an end-to-end model.

Original language	English
Title of host publication	Image and Graphics - 11th International Conference, ICIG 2021, Proceedings
Editors	Yuxin Peng, Shi-Min Hu, Moncef Gabbouj, Kun Zhou, Michael Elad, Kun Xu
Publisher	Springer Science and Business Media Deutschland GmbH
Pages	742-753
Number of pages	12
ISBN (Print)	9783030873578
DOIs	https://doi.org/10.1007/978-3-030-87358-5_60
State	Published - 2021
Event	11th International Conference on Image and Graphics, ICIG 2021 - Haikou, China Duration: 06 08 2021 → 08 08 2021

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	12889 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	11th International Conference on Image and Graphics, ICIG 2021
Country/Territory	China
City	Haikou
Period	06/08/21 → 08/08/21

Bibliographical note

Publisher Copyright:
© 2021, Springer Nature Switzerland AG.

Keywords

Deep learning
Diffusion tensor imaging
Integration
Parkinson’s disease
Regional
Sub-models

Access to Document

10.1007/978-3-030-87358-5_60

Cite this

Zhao, H., Tsai, C. C., Zhu, C., Zhou, M., Wang, J. J., & Liu, Y. (2021). Deep Learning-Based Regional Sub-models Integration for Parkinson’s Disease Diagnosis Using Diffusion Tensor Imaging. In Y. Peng, S.-M. Hu, M. Gabbouj, K. Zhou, M. Elad, & K. Xu (Eds.), Image and Graphics - 11th International Conference, ICIG 2021, Proceedings (pp. 742-753). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 12889 LNCS). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-3-030-87358-5_60

Zhao, Hengling ; Tsai, Chih Chien ; Zhu, Ce et al. / Deep Learning-Based Regional Sub-models Integration for Parkinson’s Disease Diagnosis Using Diffusion Tensor Imaging. Image and Graphics - 11th International Conference, ICIG 2021, Proceedings. editor / Yuxin Peng ; Shi-Min Hu ; Moncef Gabbouj ; Kun Zhou ; Michael Elad ; Kun Xu. Springer Science and Business Media Deutschland GmbH, 2021. pp. 742-753 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{fc459425ac73446097b27dc81ce87bd5,

title = "Deep Learning-Based Regional Sub-models Integration for Parkinson{\textquoteright}s Disease Diagnosis Using Diffusion Tensor Imaging",

abstract = "Parkinson{\textquoteright}s disease (PD) is a neurodegenerative disease. PD patients may have serious movement disorders and mental problems. The current diagnosis requires a professionally trained medical doctor to take a long period for it. Different doctors may even have different accuracies. Recently advances in deep learning-based medical image classification make it is possible to diagnose PD automatically. Different from most of the existing works on magnetic resonance images, we use diffusion tensor imaging (DTI) in that it can reflect functional data of the brain. We propose a sub-models integration framework based on convolutional neural networks (CNNs) for Parkinson{\textquoteright}s disease. Each sub-region of the brain is used to train a unique CNN model, named sub-model, and the selective stacking algorithm is used to screen these sub-models. It obtains the classification accuracy of 92.4% on the cross-validation dataset. In addition, it can provide that which sub-regions play a role in the judgment of the final result so that this framework has stronger practical application than an end-to-end model.",

keywords = "Deep learning, Diffusion tensor imaging, Integration, Parkinson{\textquoteright}s disease, Regional, Sub-models",

author = "Hengling Zhao and Tsai, {Chih Chien} and Ce Zhu and Mingyi Zhou and Wang, {Jiun Jie} and Yipeng Liu",

note = "Publisher Copyright: {\textcopyright} 2021, Springer Nature Switzerland AG.; 11th International Conference on Image and Graphics, ICIG 2021 ; Conference date: 06-08-2021 Through 08-08-2021",

year = "2021",

doi = "10.1007/978-3-030-87358-5_60",

language = "英语",

isbn = "9783030873578",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Science and Business Media Deutschland GmbH",

pages = "742--753",

editor = "Yuxin Peng and Shi-Min Hu and Moncef Gabbouj and Kun Zhou and Michael Elad and Kun Xu",

booktitle = "Image and Graphics - 11th International Conference, ICIG 2021, Proceedings",

address = "德国",

}

Zhao, H, Tsai, CC, Zhu, C, Zhou, M, Wang, JJ & Liu, Y 2021, Deep Learning-Based Regional Sub-models Integration for Parkinson’s Disease Diagnosis Using Diffusion Tensor Imaging. in Y Peng, S-M Hu, M Gabbouj, K Zhou, M Elad & K Xu (eds), Image and Graphics - 11th International Conference, ICIG 2021, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 12889 LNCS, Springer Science and Business Media Deutschland GmbH, pp. 742-753, 11th International Conference on Image and Graphics, ICIG 2021, Haikou, China, 06/08/21. https://doi.org/10.1007/978-3-030-87358-5_60

Deep Learning-Based Regional Sub-models Integration for Parkinson’s Disease Diagnosis Using Diffusion Tensor Imaging. / Zhao, Hengling; Tsai, Chih Chien; Zhu, Ce et al.
Image and Graphics - 11th International Conference, ICIG 2021, Proceedings. ed. / Yuxin Peng; Shi-Min Hu; Moncef Gabbouj; Kun Zhou; Michael Elad; Kun Xu. Springer Science and Business Media Deutschland GmbH, 2021. p. 742-753 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 12889 LNCS).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Deep Learning-Based Regional Sub-models Integration for Parkinson’s Disease Diagnosis Using Diffusion Tensor Imaging

AU - Zhao, Hengling

AU - Tsai, Chih Chien

AU - Zhu, Ce

AU - Zhou, Mingyi

AU - Wang, Jiun Jie

AU - Liu, Yipeng

PY - 2021

Y1 - 2021

N2 - Parkinson’s disease (PD) is a neurodegenerative disease. PD patients may have serious movement disorders and mental problems. The current diagnosis requires a professionally trained medical doctor to take a long period for it. Different doctors may even have different accuracies. Recently advances in deep learning-based medical image classification make it is possible to diagnose PD automatically. Different from most of the existing works on magnetic resonance images, we use diffusion tensor imaging (DTI) in that it can reflect functional data of the brain. We propose a sub-models integration framework based on convolutional neural networks (CNNs) for Parkinson’s disease. Each sub-region of the brain is used to train a unique CNN model, named sub-model, and the selective stacking algorithm is used to screen these sub-models. It obtains the classification accuracy of 92.4% on the cross-validation dataset. In addition, it can provide that which sub-regions play a role in the judgment of the final result so that this framework has stronger practical application than an end-to-end model.

AB - Parkinson’s disease (PD) is a neurodegenerative disease. PD patients may have serious movement disorders and mental problems. The current diagnosis requires a professionally trained medical doctor to take a long period for it. Different doctors may even have different accuracies. Recently advances in deep learning-based medical image classification make it is possible to diagnose PD automatically. Different from most of the existing works on magnetic resonance images, we use diffusion tensor imaging (DTI) in that it can reflect functional data of the brain. We propose a sub-models integration framework based on convolutional neural networks (CNNs) for Parkinson’s disease. Each sub-region of the brain is used to train a unique CNN model, named sub-model, and the selective stacking algorithm is used to screen these sub-models. It obtains the classification accuracy of 92.4% on the cross-validation dataset. In addition, it can provide that which sub-regions play a role in the judgment of the final result so that this framework has stronger practical application than an end-to-end model.

KW - Deep learning

KW - Diffusion tensor imaging

KW - Integration

KW - Parkinson’s disease

KW - Regional

KW - Sub-models

UR - http://www.scopus.com/inward/record.url?scp=85117068796&partnerID=8YFLogxK

U2 - 10.1007/978-3-030-87358-5_60

DO - 10.1007/978-3-030-87358-5_60

M3 - 会议稿件

AN - SCOPUS:85117068796

SN - 9783030873578

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

SP - 742

EP - 753

BT - Image and Graphics - 11th International Conference, ICIG 2021, Proceedings

A2 - Peng, Yuxin

A2 - Hu, Shi-Min

A2 - Gabbouj, Moncef

A2 - Zhou, Kun

A2 - Elad, Michael

A2 - Xu, Kun

PB - Springer Science and Business Media Deutschland GmbH

T2 - 11th International Conference on Image and Graphics, ICIG 2021

Y2 - 6 August 2021 through 8 August 2021

ER -

Zhao H, Tsai CC, Zhu C, Zhou M, Wang JJ, Liu Y. Deep Learning-Based Regional Sub-models Integration for Parkinson’s Disease Diagnosis Using Diffusion Tensor Imaging. In Peng Y, Hu SM, Gabbouj M, Zhou K, Elad M, Xu K, editors, Image and Graphics - 11th International Conference, ICIG 2021, Proceedings. Springer Science and Business Media Deutschland GmbH. 2021. p. 742-753. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-030-87358-5_60

Deep Learning-Based Regional Sub-models Integration for Parkinson’s Disease Diagnosis Using Diffusion Tensor Imaging

Abstract

Publication series

Conference

Bibliographical note

Keywords

Access to Document

Other files and links

Fingerprint

Cite this