XmoNet: a Fully Convolutional Network for Cross-Modality MR Image Inference

Sophia Bano; Muhammad Asad; Ahmed Fetit; Islem Rekik

XmoNet: a Fully Convolutional Network for Cross-Modality MR Image Inference

Sophia Bano (Lead / Corresponding author)
, Muhammad Asad
, Ahmed Fetit
, Islem Rekik

Research output: Chapter in Book/Report/Conference proceeding › Chapter (peer-reviewed) › peer-review

304 Downloads (Pure)

Abstract

Magnetic resonance imaging (MRI) can generate multimodal scans with complementary contrast information, capturing various anatomical or functional properties of organs of interest. But whilst the acquisition of multiple modalities is favourable in clinical and research settings, it is hindered by a range of practical factors that include cost and imaging artefacts.We propose XmoNet, a deep-learning architecture based on fully convolutional networks (FCNs) that enables cross-modality MR image inference. This multiple branch architecture operates on various levels of image spatial resolutions, encoding rich feature hierarchies suited for this image generation task. We illustrate the utility of XmoNet in learning the mapping between heterogeneous T1- and T2-weighted MRI scans for accurate and realistic image synthesis in a preliminary analysis. Our ndings support scaling the work to include larger samples and additional modalities.

Original language	English
Title of host publication	PRedictive Intelligence in MEdicine
Subtitle of host publication	First International Workshop, PRIME 2018, Held in Conjunction with MICCAI 2018, Granada, Spain, September 16, 2018, Proceedings
Editors	Islem Rekik, Gozde Unal, Ehsan Adeli, Sang Hyun Park
Publisher	Springer International Publishing
Number of pages	9
ISBN (Electronic)	9783030003203
ISBN (Print)	9783030003197
Publication status	Published - 2018
Event	First International Workshop, PRIME 2018, Held in Conjunction with MICCAI 2018, Granada, Spain, September 16, 2018 - Granada, Spain Duration: 16 Sept 2018 → 20 Sept 2018 https://www.miccai2018.org/en/

Publication series

Name	Lecture Notes in Computer Science
Publisher	Springer International Publishing
ISSN (Print)	0302-9743

Conference

Conference	First International Workshop, PRIME 2018, Held in Conjunction with MICCAI 2018, Granada, Spain, September 16, 2018
Country/Territory	Spain
City	Granada
Period	16/09/18 → 20/09/18
Internet address	https://www.miccai2018.org/en/

Keywords

fully convolutional networks
MRI
multimodal
image generation

Access to Document

Author Accepted ManuscriptAccepted author manuscript, 1.6 MB

https://www.springerprofessional.de/en/xmonet-a-fully-convolutional-network-for-cross-modality-mr-image/16118204

Cite this

Bano, S., Asad, M., Fetit, A., & Rekik, I. (2018). XmoNet: a Fully Convolutional Network for Cross-Modality MR Image Inference. In I. Rekik, G. Unal, E. Adeli, & S. H. Park (Eds.), PRedictive Intelligence in MEdicine: First International Workshop, PRIME 2018, Held in Conjunction with MICCAI 2018, Granada, Spain, September 16, 2018, Proceedings (Lecture Notes in Computer Science). Springer International Publishing. https://www.springerprofessional.de/en/xmonet-a-fully-convolutional-network-for-cross-modality-mr-image/16118204

Bano, Sophia ; Asad, Muhammad ; Fetit, Ahmed et al. / XmoNet : a Fully Convolutional Network for Cross-Modality MR Image Inference. PRedictive Intelligence in MEdicine: First International Workshop, PRIME 2018, Held in Conjunction with MICCAI 2018, Granada, Spain, September 16, 2018, Proceedings. editor / Islem Rekik ; Gozde Unal ; Ehsan Adeli ; Sang Hyun Park. Springer International Publishing, 2018. (Lecture Notes in Computer Science).

@inbook{f04883fac19742fdbf622fc733901d64,

title = "XmoNet: a Fully Convolutional Network for Cross-Modality MR Image Inference",

abstract = "Magnetic resonance imaging (MRI) can generate multimodal scans with complementary contrast information, capturing various anatomical or functional properties of organs of interest. But whilst the acquisition of multiple modalities is favourable in clinical and research settings, it is hindered by a range of practical factors that include cost and imaging artefacts.We propose XmoNet, a deep-learning architecture based on fully convolutional networks (FCNs) that enables cross-modality MR image inference. This multiple branch architecture operates on various levels of image spatial resolutions, encoding rich feature hierarchies suited for this image generation task. We illustrate the utility of XmoNet in learning the mapping between heterogeneous T1- and T2-weighted MRI scans for accurate and realistic image synthesis in a preliminary analysis. Our ndings support scaling the work to include larger samples and additional modalities.",

keywords = "fully convolutional networks, MRI, multimodal, image generation",

author = "Sophia Bano and Muhammad Asad and Ahmed Fetit and Islem Rekik",

year = "2018",

language = "English",

isbn = "9783030003197",

series = "Lecture Notes in Computer Science",

publisher = "Springer International Publishing",

editor = "Islem Rekik and Gozde Unal and Ehsan Adeli and Park, \{Sang Hyun\}",

booktitle = "PRedictive Intelligence in MEdicine",

address = "Switzerland",

note = "First International Workshop, PRIME 2018, Held in Conjunction with MICCAI 2018, Granada, Spain, September 16, 2018 ; Conference date: 16-09-2018 Through 20-09-2018",

url = "https://www.miccai2018.org/en/",

}

Bano, S, Asad, M, Fetit, A & Rekik, I 2018, XmoNet: a Fully Convolutional Network for Cross-Modality MR Image Inference. in I Rekik, G Unal, E Adeli & SH Park (eds), PRedictive Intelligence in MEdicine: First International Workshop, PRIME 2018, Held in Conjunction with MICCAI 2018, Granada, Spain, September 16, 2018, Proceedings. Lecture Notes in Computer Science, Springer International Publishing, First International Workshop, PRIME 2018, Held in Conjunction with MICCAI 2018, Granada, Spain, September 16, 2018, Granada, Spain, 16/09/18. <https://www.springerprofessional.de/en/xmonet-a-fully-convolutional-network-for-cross-modality-mr-image/16118204>

XmoNet: a Fully Convolutional Network for Cross-Modality MR Image Inference. / Bano, Sophia (Lead / Corresponding author); Asad, Muhammad; Fetit, Ahmed et al.
PRedictive Intelligence in MEdicine: First International Workshop, PRIME 2018, Held in Conjunction with MICCAI 2018, Granada, Spain, September 16, 2018, Proceedings. ed. / Islem Rekik; Gozde Unal; Ehsan Adeli; Sang Hyun Park. Springer International Publishing, 2018. (Lecture Notes in Computer Science).

Research output: Chapter in Book/Report/Conference proceeding › Chapter (peer-reviewed) › peer-review

TY - CHAP

T1 - XmoNet

T2 - First International Workshop, PRIME 2018, Held in Conjunction with MICCAI 2018, Granada, Spain, September 16, 2018

AU - Bano, Sophia

AU - Asad, Muhammad

AU - Fetit, Ahmed

AU - Rekik, Islem

PY - 2018

Y1 - 2018

N2 - Magnetic resonance imaging (MRI) can generate multimodal scans with complementary contrast information, capturing various anatomical or functional properties of organs of interest. But whilst the acquisition of multiple modalities is favourable in clinical and research settings, it is hindered by a range of practical factors that include cost and imaging artefacts.We propose XmoNet, a deep-learning architecture based on fully convolutional networks (FCNs) that enables cross-modality MR image inference. This multiple branch architecture operates on various levels of image spatial resolutions, encoding rich feature hierarchies suited for this image generation task. We illustrate the utility of XmoNet in learning the mapping between heterogeneous T1- and T2-weighted MRI scans for accurate and realistic image synthesis in a preliminary analysis. Our ndings support scaling the work to include larger samples and additional modalities.

AB - Magnetic resonance imaging (MRI) can generate multimodal scans with complementary contrast information, capturing various anatomical or functional properties of organs of interest. But whilst the acquisition of multiple modalities is favourable in clinical and research settings, it is hindered by a range of practical factors that include cost and imaging artefacts.We propose XmoNet, a deep-learning architecture based on fully convolutional networks (FCNs) that enables cross-modality MR image inference. This multiple branch architecture operates on various levels of image spatial resolutions, encoding rich feature hierarchies suited for this image generation task. We illustrate the utility of XmoNet in learning the mapping between heterogeneous T1- and T2-weighted MRI scans for accurate and realistic image synthesis in a preliminary analysis. Our ndings support scaling the work to include larger samples and additional modalities.

KW - fully convolutional networks

KW - MRI

KW - multimodal

KW - image generation

M3 - Chapter (peer-reviewed)

SN - 9783030003197

T3 - Lecture Notes in Computer Science

BT - PRedictive Intelligence in MEdicine

A2 - Rekik, Islem

A2 - Unal, Gozde

A2 - Adeli, Ehsan

A2 - Park, Sang Hyun

PB - Springer International Publishing

Y2 - 16 September 2018 through 20 September 2018

ER -

Bano S, Asad M, Fetit A, Rekik I. XmoNet: a Fully Convolutional Network for Cross-Modality MR Image Inference. In Rekik I, Unal G, Adeli E, Park SH, editors, PRedictive Intelligence in MEdicine: First International Workshop, PRIME 2018, Held in Conjunction with MICCAI 2018, Granada, Spain, September 16, 2018, Proceedings. Springer International Publishing. 2018. (Lecture Notes in Computer Science).

XmoNet: a Fully Convolutional Network for Cross-Modality MR Image Inference

Abstract

Publication series

Conference

Keywords

Access to Document

Fingerprint

Cite this