A supervoxel based random forest synthesis framework for bidirectional MR/CT synthesis

Can Zhao, Aaron Carass, Junghoon Lee, Amod Jog, Jerry Ladd Prince

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

Abstract

Synthesizing magnetic resonance (MR) and computed tomography (CT) images (from each other) has important implications for clinical neuroimaging. The MR to CT direction is critical for MRI-based radiotherapy planning and dose computation, whereas the CT to MR direction can provide an economic alternative to real MRI for image processing tasks. Additionally, synthesis in both directions can enhance MR/CT multi-modal image registration. Existing approaches have focused on synthesizing CT from MR. In this paper, we propose a multi-atlas based hybrid method to synthesize T1-weighted MR images from CT and CT images from T1-weighted MR images using a common framework. The task is carried out by: (a) computing a label field based on supervoxels for the subject image using joint label fusion; (b) correcting this result using a random forest classifier (RF-C); (c) spatial smoothing using a Markov random field; (d) synthesizing intensities using a set of RF regressors, one trained for each label. The algorithm is evaluated using a set of six registered CT and MR image pairs of the whole head.

Original languageEnglish (US)
Title of host publicationSimulation and Synthesis in Medical Imaging - 2nd International Workshop, SASHIMI 2017, Held in Conjunction with MICCAI 2017, Proceedings
PublisherSpringer Verlag
Pages33-40
Number of pages8
ISBN (Print)9783319681269
DOIs
StatePublished - Jan 1 2017
Event2nd International Workshop on Simulation and Synthesis in Medical Imaging, SASHIMI 2017 Held in Conjunction with the 20th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2017 - Quebec City, Canada
Duration: Sep 10 2017Sep 10 2017

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume10557 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Other

Other2nd International Workshop on Simulation and Synthesis in Medical Imaging, SASHIMI 2017 Held in Conjunction with the 20th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2017
CountryCanada
CityQuebec City
Period9/10/179/10/17

Fingerprint

Magnetic Resonance
Random Forest
Computed Tomography
Magnetic resonance
Tomography
Synthesis
Magnetic Resonance Image
Labels
Magnetic resonance imaging
Neuroimaging
Radiotherapy
Framework
Atlas
Image registration
Image Registration
Hybrid Method
Random Field
Dosimetry
Smoothing
Image Processing

Keywords

  • CT
  • JLF
  • MR
  • MRF
  • Random forest
  • Segmentation
  • Synthesis

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Cite this

Zhao, C., Carass, A., Lee, J., Jog, A., & Prince, J. L. (2017). A supervoxel based random forest synthesis framework for bidirectional MR/CT synthesis. In Simulation and Synthesis in Medical Imaging - 2nd International Workshop, SASHIMI 2017, Held in Conjunction with MICCAI 2017, Proceedings (pp. 33-40). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 10557 LNCS). Springer Verlag. https://doi.org/10.1007/978-3-319-68127-6_4

A supervoxel based random forest synthesis framework for bidirectional MR/CT synthesis. / Zhao, Can; Carass, Aaron; Lee, Junghoon; Jog, Amod; Prince, Jerry Ladd.

Simulation and Synthesis in Medical Imaging - 2nd International Workshop, SASHIMI 2017, Held in Conjunction with MICCAI 2017, Proceedings. Springer Verlag, 2017. p. 33-40 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 10557 LNCS).

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

Zhao, C, Carass, A, Lee, J, Jog, A & Prince, JL 2017, A supervoxel based random forest synthesis framework for bidirectional MR/CT synthesis. in Simulation and Synthesis in Medical Imaging - 2nd International Workshop, SASHIMI 2017, Held in Conjunction with MICCAI 2017, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 10557 LNCS, Springer Verlag, pp. 33-40, 2nd International Workshop on Simulation and Synthesis in Medical Imaging, SASHIMI 2017 Held in Conjunction with the 20th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2017, Quebec City, Canada, 9/10/17. https://doi.org/10.1007/978-3-319-68127-6_4
Zhao C, Carass A, Lee J, Jog A, Prince JL. A supervoxel based random forest synthesis framework for bidirectional MR/CT synthesis. In Simulation and Synthesis in Medical Imaging - 2nd International Workshop, SASHIMI 2017, Held in Conjunction with MICCAI 2017, Proceedings. Springer Verlag. 2017. p. 33-40. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-319-68127-6_4
Zhao, Can ; Carass, Aaron ; Lee, Junghoon ; Jog, Amod ; Prince, Jerry Ladd. / A supervoxel based random forest synthesis framework for bidirectional MR/CT synthesis. Simulation and Synthesis in Medical Imaging - 2nd International Workshop, SASHIMI 2017, Held in Conjunction with MICCAI 2017, Proceedings. Springer Verlag, 2017. pp. 33-40 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{a575106c6fa449018788eecc717a0dd6,
title = "A supervoxel based random forest synthesis framework for bidirectional MR/CT synthesis",
abstract = "Synthesizing magnetic resonance (MR) and computed tomography (CT) images (from each other) has important implications for clinical neuroimaging. The MR to CT direction is critical for MRI-based radiotherapy planning and dose computation, whereas the CT to MR direction can provide an economic alternative to real MRI for image processing tasks. Additionally, synthesis in both directions can enhance MR/CT multi-modal image registration. Existing approaches have focused on synthesizing CT from MR. In this paper, we propose a multi-atlas based hybrid method to synthesize T1-weighted MR images from CT and CT images from T1-weighted MR images using a common framework. The task is carried out by: (a) computing a label field based on supervoxels for the subject image using joint label fusion; (b) correcting this result using a random forest classifier (RF-C); (c) spatial smoothing using a Markov random field; (d) synthesizing intensities using a set of RF regressors, one trained for each label. The algorithm is evaluated using a set of six registered CT and MR image pairs of the whole head.",
keywords = "CT, JLF, MR, MRF, Random forest, Segmentation, Synthesis",
author = "Can Zhao and Aaron Carass and Junghoon Lee and Amod Jog and Prince, {Jerry Ladd}",
year = "2017",
month = "1",
day = "1",
doi = "10.1007/978-3-319-68127-6_4",
language = "English (US)",
isbn = "9783319681269",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "33--40",
booktitle = "Simulation and Synthesis in Medical Imaging - 2nd International Workshop, SASHIMI 2017, Held in Conjunction with MICCAI 2017, Proceedings",

}

TY - GEN

T1 - A supervoxel based random forest synthesis framework for bidirectional MR/CT synthesis

AU - Zhao, Can

AU - Carass, Aaron

AU - Lee, Junghoon

AU - Jog, Amod

AU - Prince, Jerry Ladd

PY - 2017/1/1

Y1 - 2017/1/1

N2 - Synthesizing magnetic resonance (MR) and computed tomography (CT) images (from each other) has important implications for clinical neuroimaging. The MR to CT direction is critical for MRI-based radiotherapy planning and dose computation, whereas the CT to MR direction can provide an economic alternative to real MRI for image processing tasks. Additionally, synthesis in both directions can enhance MR/CT multi-modal image registration. Existing approaches have focused on synthesizing CT from MR. In this paper, we propose a multi-atlas based hybrid method to synthesize T1-weighted MR images from CT and CT images from T1-weighted MR images using a common framework. The task is carried out by: (a) computing a label field based on supervoxels for the subject image using joint label fusion; (b) correcting this result using a random forest classifier (RF-C); (c) spatial smoothing using a Markov random field; (d) synthesizing intensities using a set of RF regressors, one trained for each label. The algorithm is evaluated using a set of six registered CT and MR image pairs of the whole head.

AB - Synthesizing magnetic resonance (MR) and computed tomography (CT) images (from each other) has important implications for clinical neuroimaging. The MR to CT direction is critical for MRI-based radiotherapy planning and dose computation, whereas the CT to MR direction can provide an economic alternative to real MRI for image processing tasks. Additionally, synthesis in both directions can enhance MR/CT multi-modal image registration. Existing approaches have focused on synthesizing CT from MR. In this paper, we propose a multi-atlas based hybrid method to synthesize T1-weighted MR images from CT and CT images from T1-weighted MR images using a common framework. The task is carried out by: (a) computing a label field based on supervoxels for the subject image using joint label fusion; (b) correcting this result using a random forest classifier (RF-C); (c) spatial smoothing using a Markov random field; (d) synthesizing intensities using a set of RF regressors, one trained for each label. The algorithm is evaluated using a set of six registered CT and MR image pairs of the whole head.

KW - CT

KW - JLF

KW - MR

KW - MRF

KW - Random forest

KW - Segmentation

KW - Synthesis

UR - http://www.scopus.com/inward/record.url?scp=85031402661&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85031402661&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-68127-6_4

DO - 10.1007/978-3-319-68127-6_4

M3 - Conference contribution

C2 - 30221260

AN - SCOPUS:85031402661

SN - 9783319681269

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

SP - 33

EP - 40

BT - Simulation and Synthesis in Medical Imaging - 2nd International Workshop, SASHIMI 2017, Held in Conjunction with MICCAI 2017, Proceedings

PB - Springer Verlag

ER -