Deep MR Fingerprinting with total-variation and low-rank subspace priors

Mohammad Golbabaee, Carolin M. Pirkl, Marion I. Menzel, Guido Buonincontri, Pedro A. Gómez

Research output: Contribution to conferencePaper

Abstract

Deep learning (DL) has recently emerged to address the heavy storage and computation requirements of the baseline dictionary-matching (DM) for Magnetic Resonance Fingerprinting (MRF) reconstruction. Fed with non-iterated back-projected images, the network is unable to fully resolve spatially-correlated corruptions caused from the undersampling artefacts. We propose an accelerated iterative reconstruction to minimize these artefacts before feeding into the network. This is done through a convex regularization that jointly promotes spatio-temporal regularities of the MRF time-series. Except for training, the rest of the parameter estimation pipeline is dictionary-free. We validate the proposed approach on synthetic and in-vivo datasets.
Original languageEnglish
Publication statusPublished - 16 May 2019
EventISMRM 27th Annual Meeting and Exhibition - Montreal, Canada
Duration: 11 May 201916 May 2019
https://www.ismrm.org/19m/

Conference

ConferenceISMRM 27th Annual Meeting and Exhibition
Abbreviated titleISMRM 2019
CountryCanada
CityMontreal
Period11/05/1916/05/19
Internet address

Cite this

Golbabaee, M., Pirkl, C. M., Menzel, M. I., Buonincontri, G., & Gómez, P. A. (2019). Deep MR Fingerprinting with total-variation and low-rank subspace priors. Paper presented at ISMRM 27th Annual Meeting and Exhibition, Montreal, Canada.

Deep MR Fingerprinting with total-variation and low-rank subspace priors. / Golbabaee, Mohammad; Pirkl, Carolin M. ; Menzel, Marion I.; Buonincontri, Guido; Gómez, Pedro A. .

2019. Paper presented at ISMRM 27th Annual Meeting and Exhibition, Montreal, Canada.

Research output: Contribution to conferencePaper

Golbabaee, M, Pirkl, CM, Menzel, MI, Buonincontri, G & Gómez, PA 2019, 'Deep MR Fingerprinting with total-variation and low-rank subspace priors' Paper presented at ISMRM 27th Annual Meeting and Exhibition, Montreal, Canada, 11/05/19 - 16/05/19, .
Golbabaee M, Pirkl CM, Menzel MI, Buonincontri G, Gómez PA. Deep MR Fingerprinting with total-variation and low-rank subspace priors. 2019. Paper presented at ISMRM 27th Annual Meeting and Exhibition, Montreal, Canada.
Golbabaee, Mohammad ; Pirkl, Carolin M. ; Menzel, Marion I. ; Buonincontri, Guido ; Gómez, Pedro A. . / Deep MR Fingerprinting with total-variation and low-rank subspace priors. Paper presented at ISMRM 27th Annual Meeting and Exhibition, Montreal, Canada.
@conference{6cee97393d294e40a90880be1be010ce,
title = "Deep MR Fingerprinting with total-variation and low-rank subspace priors",
abstract = "Deep learning (DL) has recently emerged to address the heavy storage and computation requirements of the baseline dictionary-matching (DM) for Magnetic Resonance Fingerprinting (MRF) reconstruction. Fed with non-iterated back-projected images, the network is unable to fully resolve spatially-correlated corruptions caused from the undersampling artefacts. We propose an accelerated iterative reconstruction to minimize these artefacts before feeding into the network. This is done through a convex regularization that jointly promotes spatio-temporal regularities of the MRF time-series. Except for training, the rest of the parameter estimation pipeline is dictionary-free. We validate the proposed approach on synthetic and in-vivo datasets.",
author = "Mohammad Golbabaee and Pirkl, {Carolin M.} and Menzel, {Marion I.} and Guido Buonincontri and G{\'o}mez, {Pedro A.}",
year = "2019",
month = "5",
day = "16",
language = "English",
note = "ISMRM 27th Annual Meeting and Exhibition, ISMRM 2019 ; Conference date: 11-05-2019 Through 16-05-2019",
url = "https://www.ismrm.org/19m/",

}

TY - CONF

T1 - Deep MR Fingerprinting with total-variation and low-rank subspace priors

AU - Golbabaee, Mohammad

AU - Pirkl, Carolin M.

AU - Menzel, Marion I.

AU - Buonincontri, Guido

AU - Gómez, Pedro A.

PY - 2019/5/16

Y1 - 2019/5/16

N2 - Deep learning (DL) has recently emerged to address the heavy storage and computation requirements of the baseline dictionary-matching (DM) for Magnetic Resonance Fingerprinting (MRF) reconstruction. Fed with non-iterated back-projected images, the network is unable to fully resolve spatially-correlated corruptions caused from the undersampling artefacts. We propose an accelerated iterative reconstruction to minimize these artefacts before feeding into the network. This is done through a convex regularization that jointly promotes spatio-temporal regularities of the MRF time-series. Except for training, the rest of the parameter estimation pipeline is dictionary-free. We validate the proposed approach on synthetic and in-vivo datasets.

AB - Deep learning (DL) has recently emerged to address the heavy storage and computation requirements of the baseline dictionary-matching (DM) for Magnetic Resonance Fingerprinting (MRF) reconstruction. Fed with non-iterated back-projected images, the network is unable to fully resolve spatially-correlated corruptions caused from the undersampling artefacts. We propose an accelerated iterative reconstruction to minimize these artefacts before feeding into the network. This is done through a convex regularization that jointly promotes spatio-temporal regularities of the MRF time-series. Except for training, the rest of the parameter estimation pipeline is dictionary-free. We validate the proposed approach on synthetic and in-vivo datasets.

M3 - Paper

ER -