Joint intensity fusion image synthesis applied to multiple sclerosis lesion segmentation

Greg M. Fleishman, Alessandra Valcarcel, Dzung L. Pham, Snehashis Roy, Peter Calabresi, Paul Yushkevich, Russell T. Shinohara, Ipek Oguz

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

Abstract

We propose a new approach to Multiple Sclerosis lesion segmentation that utilizes synthesized images. A new method of image synthesis is considered: joint intensity fusion (JIF). JIF synthesizes an image from a library of deformably registered and intensity normalized atlases. Each location in the synthesized image is a weighted average of the registered atlases; atlas weights vary spatially. The weights are determined using the joint label fusion (JLF) framework. The primary methodological contribution is the application of JLF to MRI signal directly rather than labels. Synthesized images are then used as additional features in a lesion segmentation task using the OASIS classifier, a logistic regression model on intensities from multiple modalities. The addition of JIF synthesized images improved the Dice-Sorensen coefficient (relative to manually drawn gold standards) of lesion segmentations over the standard model segmentations by 0.0462 ± 0.0050 (mean ± standard deviation) at optimal threshold over all subjects and 10 separate training/testing folds.

Original languageEnglish (US)
Title of host publicationBrainlesion
Subtitle of host publicationGlioma, Multiple Sclerosis, Stroke and Traumatic Brain Injuries - 3rd International Workshop, BrainLes 2017, Held in Conjunction with MICCAI 2017, Revised Selected Papers
PublisherSpringer Verlag
Pages43-54
Number of pages12
ISBN (Print)9783319752372
DOIs
StatePublished - Jan 1 2018
Event3rd International Workshop on Brainlesion, BrainLes 2017 Held in Conjunction with Medical Image Computing for Computer Assisted Intervention , MICCAI 2017 - Quebec City, Canada
Duration: Sep 14 2017Sep 14 2017

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume10670 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Other

Other3rd International Workshop on Brainlesion, BrainLes 2017 Held in Conjunction with Medical Image Computing for Computer Assisted Intervention , MICCAI 2017
CountryCanada
CityQuebec City
Period9/14/179/14/17

Fingerprint

Multiple Sclerosis
Image fusion
Image Fusion
Labels
Segmentation
Synthesis
Atlas
Fusion
Magnetic resonance imaging
Logistics
Classifiers
Dice
Method of Images
Logistic Regression Model
Weighted Average
Gold
Standard deviation
Modality
Testing
Standard Model

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Cite this

Fleishman, G. M., Valcarcel, A., Pham, D. L., Roy, S., Calabresi, P., Yushkevich, P., ... Oguz, I. (2018). Joint intensity fusion image synthesis applied to multiple sclerosis lesion segmentation. In Brainlesion: Glioma, Multiple Sclerosis, Stroke and Traumatic Brain Injuries - 3rd International Workshop, BrainLes 2017, Held in Conjunction with MICCAI 2017, Revised Selected Papers (pp. 43-54). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 10670 LNCS). Springer Verlag. https://doi.org/10.1007/978-3-319-75238-9_4

Joint intensity fusion image synthesis applied to multiple sclerosis lesion segmentation. / Fleishman, Greg M.; Valcarcel, Alessandra; Pham, Dzung L.; Roy, Snehashis; Calabresi, Peter; Yushkevich, Paul; Shinohara, Russell T.; Oguz, Ipek.

Brainlesion: Glioma, Multiple Sclerosis, Stroke and Traumatic Brain Injuries - 3rd International Workshop, BrainLes 2017, Held in Conjunction with MICCAI 2017, Revised Selected Papers. Springer Verlag, 2018. p. 43-54 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 10670 LNCS).

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

Fleishman, GM, Valcarcel, A, Pham, DL, Roy, S, Calabresi, P, Yushkevich, P, Shinohara, RT & Oguz, I 2018, Joint intensity fusion image synthesis applied to multiple sclerosis lesion segmentation. in Brainlesion: Glioma, Multiple Sclerosis, Stroke and Traumatic Brain Injuries - 3rd International Workshop, BrainLes 2017, Held in Conjunction with MICCAI 2017, Revised Selected Papers. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 10670 LNCS, Springer Verlag, pp. 43-54, 3rd International Workshop on Brainlesion, BrainLes 2017 Held in Conjunction with Medical Image Computing for Computer Assisted Intervention , MICCAI 2017, Quebec City, Canada, 9/14/17. https://doi.org/10.1007/978-3-319-75238-9_4
Fleishman GM, Valcarcel A, Pham DL, Roy S, Calabresi P, Yushkevich P et al. Joint intensity fusion image synthesis applied to multiple sclerosis lesion segmentation. In Brainlesion: Glioma, Multiple Sclerosis, Stroke and Traumatic Brain Injuries - 3rd International Workshop, BrainLes 2017, Held in Conjunction with MICCAI 2017, Revised Selected Papers. Springer Verlag. 2018. p. 43-54. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-319-75238-9_4
Fleishman, Greg M. ; Valcarcel, Alessandra ; Pham, Dzung L. ; Roy, Snehashis ; Calabresi, Peter ; Yushkevich, Paul ; Shinohara, Russell T. ; Oguz, Ipek. / Joint intensity fusion image synthesis applied to multiple sclerosis lesion segmentation. Brainlesion: Glioma, Multiple Sclerosis, Stroke and Traumatic Brain Injuries - 3rd International Workshop, BrainLes 2017, Held in Conjunction with MICCAI 2017, Revised Selected Papers. Springer Verlag, 2018. pp. 43-54 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{635cd914579143b6902aff54cc92a3fe,
title = "Joint intensity fusion image synthesis applied to multiple sclerosis lesion segmentation",
abstract = "We propose a new approach to Multiple Sclerosis lesion segmentation that utilizes synthesized images. A new method of image synthesis is considered: joint intensity fusion (JIF). JIF synthesizes an image from a library of deformably registered and intensity normalized atlases. Each location in the synthesized image is a weighted average of the registered atlases; atlas weights vary spatially. The weights are determined using the joint label fusion (JLF) framework. The primary methodological contribution is the application of JLF to MRI signal directly rather than labels. Synthesized images are then used as additional features in a lesion segmentation task using the OASIS classifier, a logistic regression model on intensities from multiple modalities. The addition of JIF synthesized images improved the Dice-Sorensen coefficient (relative to manually drawn gold standards) of lesion segmentations over the standard model segmentations by 0.0462 ± 0.0050 (mean ± standard deviation) at optimal threshold over all subjects and 10 separate training/testing folds.",
author = "Fleishman, {Greg M.} and Alessandra Valcarcel and Pham, {Dzung L.} and Snehashis Roy and Peter Calabresi and Paul Yushkevich and Shinohara, {Russell T.} and Ipek Oguz",
year = "2018",
month = "1",
day = "1",
doi = "10.1007/978-3-319-75238-9_4",
language = "English (US)",
isbn = "9783319752372",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "43--54",
booktitle = "Brainlesion",

}

TY - GEN

T1 - Joint intensity fusion image synthesis applied to multiple sclerosis lesion segmentation

AU - Fleishman, Greg M.

AU - Valcarcel, Alessandra

AU - Pham, Dzung L.

AU - Roy, Snehashis

AU - Calabresi, Peter

AU - Yushkevich, Paul

AU - Shinohara, Russell T.

AU - Oguz, Ipek

PY - 2018/1/1

Y1 - 2018/1/1

N2 - We propose a new approach to Multiple Sclerosis lesion segmentation that utilizes synthesized images. A new method of image synthesis is considered: joint intensity fusion (JIF). JIF synthesizes an image from a library of deformably registered and intensity normalized atlases. Each location in the synthesized image is a weighted average of the registered atlases; atlas weights vary spatially. The weights are determined using the joint label fusion (JLF) framework. The primary methodological contribution is the application of JLF to MRI signal directly rather than labels. Synthesized images are then used as additional features in a lesion segmentation task using the OASIS classifier, a logistic regression model on intensities from multiple modalities. The addition of JIF synthesized images improved the Dice-Sorensen coefficient (relative to manually drawn gold standards) of lesion segmentations over the standard model segmentations by 0.0462 ± 0.0050 (mean ± standard deviation) at optimal threshold over all subjects and 10 separate training/testing folds.

AB - We propose a new approach to Multiple Sclerosis lesion segmentation that utilizes synthesized images. A new method of image synthesis is considered: joint intensity fusion (JIF). JIF synthesizes an image from a library of deformably registered and intensity normalized atlases. Each location in the synthesized image is a weighted average of the registered atlases; atlas weights vary spatially. The weights are determined using the joint label fusion (JLF) framework. The primary methodological contribution is the application of JLF to MRI signal directly rather than labels. Synthesized images are then used as additional features in a lesion segmentation task using the OASIS classifier, a logistic regression model on intensities from multiple modalities. The addition of JIF synthesized images improved the Dice-Sorensen coefficient (relative to manually drawn gold standards) of lesion segmentations over the standard model segmentations by 0.0462 ± 0.0050 (mean ± standard deviation) at optimal threshold over all subjects and 10 separate training/testing folds.

UR - http://www.scopus.com/inward/record.url?scp=85042541077&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85042541077&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-75238-9_4

DO - 10.1007/978-3-319-75238-9_4

M3 - Conference contribution

AN - SCOPUS:85042541077

SN - 9783319752372

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

SP - 43

EP - 54

BT - Brainlesion

PB - Springer Verlag

ER -