Deep learning based imaging data completion for improved brain disease diagnosis

Rongjian Li, Wenlu Zhang, Heung-Il Suk, Li Wang, Jiang Li, Dinggang Shen, Shuiwang Ji

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

153 Citations (Scopus)

Abstract

Combining multi-modality brain data for disease diagnosis commonly leads to improved performance. A challenge in using multi-modality data is that the data are commonly incomplete; namely, some modality might be missing for some subjects. In this work, we proposed a deep learning based framework for estimating multi-modality imaging data. Our method takes the form of convolutional neural networks, where the input and output are two volumetric modalities. The network contains a large number of trainable parameters that capture the relationship between input and output modalities. When trained on subjects with all modalities, the network can estimate the output modality given the input modality. We evaluated our method on the Alzheimer's Disease Neuroimaging Initiative (ADNI) database, where the input and output modalities are MRI and PET images, respectively. Results showed that our method significantly outperformed prior methods.

Original languageEnglish
Title of host publicationLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
PublisherSpringer Verlag
Pages305-312
Number of pages8
Volume8675 LNCS
EditionPART 3
ISBN (Print)9783319104423
DOIs
Publication statusPublished - 2014 Jan 1
Externally publishedYes
Event17th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2014 - Boston, MA, United States
Duration: 2014 Sep 142014 Sep 18

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
NumberPART 3
Volume8675 LNCS
ISSN (Print)03029743
ISSN (Electronic)16113349

Other

Other17th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2014
CountryUnited States
CityBoston, MA
Period14/9/1414/9/18

Fingerprint

Modality
Completion
Brain
Imaging
Neuroimaging
Imaging techniques
Multimodality
Magnetic resonance imaging
Neural networks
Output
Alzheimer's Disease
Learning
Deep learning
Neural Networks
Estimate

ASJC Scopus subject areas

  • Computer Science(all)
  • Theoretical Computer Science

Cite this

Li, R., Zhang, W., Suk, H-I., Wang, L., Li, J., Shen, D., & Ji, S. (2014). Deep learning based imaging data completion for improved brain disease diagnosis. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (PART 3 ed., Vol. 8675 LNCS, pp. 305-312). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 8675 LNCS, No. PART 3). Springer Verlag. https://doi.org/10.1007/978-3-319-10443-0_39

Deep learning based imaging data completion for improved brain disease diagnosis. / Li, Rongjian; Zhang, Wenlu; Suk, Heung-Il; Wang, Li; Li, Jiang; Shen, Dinggang; Ji, Shuiwang.

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 8675 LNCS PART 3. ed. Springer Verlag, 2014. p. 305-312 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 8675 LNCS, No. PART 3).

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

Li, R, Zhang, W, Suk, H-I, Wang, L, Li, J, Shen, D & Ji, S 2014, Deep learning based imaging data completion for improved brain disease diagnosis. in Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). PART 3 edn, vol. 8675 LNCS, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), no. PART 3, vol. 8675 LNCS, Springer Verlag, pp. 305-312, 17th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2014, Boston, MA, United States, 14/9/14. https://doi.org/10.1007/978-3-319-10443-0_39
Li R, Zhang W, Suk H-I, Wang L, Li J, Shen D et al. Deep learning based imaging data completion for improved brain disease diagnosis. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). PART 3 ed. Vol. 8675 LNCS. Springer Verlag. 2014. p. 305-312. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); PART 3). https://doi.org/10.1007/978-3-319-10443-0_39
Li, Rongjian ; Zhang, Wenlu ; Suk, Heung-Il ; Wang, Li ; Li, Jiang ; Shen, Dinggang ; Ji, Shuiwang. / Deep learning based imaging data completion for improved brain disease diagnosis. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 8675 LNCS PART 3. ed. Springer Verlag, 2014. pp. 305-312 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); PART 3).
@inproceedings{491cadd3f8074646b2d1ba9c7d5409ae,
title = "Deep learning based imaging data completion for improved brain disease diagnosis",
abstract = "Combining multi-modality brain data for disease diagnosis commonly leads to improved performance. A challenge in using multi-modality data is that the data are commonly incomplete; namely, some modality might be missing for some subjects. In this work, we proposed a deep learning based framework for estimating multi-modality imaging data. Our method takes the form of convolutional neural networks, where the input and output are two volumetric modalities. The network contains a large number of trainable parameters that capture the relationship between input and output modalities. When trained on subjects with all modalities, the network can estimate the output modality given the input modality. We evaluated our method on the Alzheimer's Disease Neuroimaging Initiative (ADNI) database, where the input and output modalities are MRI and PET images, respectively. Results showed that our method significantly outperformed prior methods.",
author = "Rongjian Li and Wenlu Zhang and Heung-Il Suk and Li Wang and Jiang Li and Dinggang Shen and Shuiwang Ji",
year = "2014",
month = "1",
day = "1",
doi = "10.1007/978-3-319-10443-0_39",
language = "English",
isbn = "9783319104423",
volume = "8675 LNCS",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
number = "PART 3",
pages = "305--312",
booktitle = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
edition = "PART 3",

}

TY - GEN

T1 - Deep learning based imaging data completion for improved brain disease diagnosis

AU - Li, Rongjian

AU - Zhang, Wenlu

AU - Suk, Heung-Il

AU - Wang, Li

AU - Li, Jiang

AU - Shen, Dinggang

AU - Ji, Shuiwang

PY - 2014/1/1

Y1 - 2014/1/1

N2 - Combining multi-modality brain data for disease diagnosis commonly leads to improved performance. A challenge in using multi-modality data is that the data are commonly incomplete; namely, some modality might be missing for some subjects. In this work, we proposed a deep learning based framework for estimating multi-modality imaging data. Our method takes the form of convolutional neural networks, where the input and output are two volumetric modalities. The network contains a large number of trainable parameters that capture the relationship between input and output modalities. When trained on subjects with all modalities, the network can estimate the output modality given the input modality. We evaluated our method on the Alzheimer's Disease Neuroimaging Initiative (ADNI) database, where the input and output modalities are MRI and PET images, respectively. Results showed that our method significantly outperformed prior methods.

AB - Combining multi-modality brain data for disease diagnosis commonly leads to improved performance. A challenge in using multi-modality data is that the data are commonly incomplete; namely, some modality might be missing for some subjects. In this work, we proposed a deep learning based framework for estimating multi-modality imaging data. Our method takes the form of convolutional neural networks, where the input and output are two volumetric modalities. The network contains a large number of trainable parameters that capture the relationship between input and output modalities. When trained on subjects with all modalities, the network can estimate the output modality given the input modality. We evaluated our method on the Alzheimer's Disease Neuroimaging Initiative (ADNI) database, where the input and output modalities are MRI and PET images, respectively. Results showed that our method significantly outperformed prior methods.

UR - http://www.scopus.com/inward/record.url?scp=84906979740&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84906979740&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-10443-0_39

DO - 10.1007/978-3-319-10443-0_39

M3 - Conference contribution

C2 - 25320813

AN - SCOPUS:84906979740

SN - 9783319104423

VL - 8675 LNCS

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

SP - 305

EP - 312

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

PB - Springer Verlag

ER -