Header

UZH-Logo

Maintenance Infos

SRflow: Deep learning based super-resolution of 4D-flow MRI data


Shit, Suprosanna; Zimmermann, Judith; Ezhov, Ivan; Paetzold, Johannes C; Sanches, Augusto F; Pirkl, Carolin; Menze, Bjoern H (2022). SRflow: Deep learning based super-resolution of 4D-flow MRI data. Frontiers in Artificial Intelligence, 5:928181.

Abstract

Exploiting 4D-flow magnetic resonance imaging (MRI) data to quantify hemodynamics requires an adequate spatio-temporal vector field resolution at a low noise level. To address this challenge, we provide a learned solution to super-resolve in vivo 4D-flow MRI data at a post-processing level. We propose a deep convolutional neural network (CNN) that learns the inter-scale relationship of the velocity vector map and leverages an efficient residual learning scheme to make it computationally feasible. A novel, direction-sensitive, and robust loss function is crucial to learning vector-field data. We present a detailed comparative study between the proposed super-resolution and the conventional cubic B-spline based vector-field super-resolution. Our method improves the peak-velocity to noise ratio of the flow field by 10 and 30% for in vivo cardiovascular and cerebrovascular data, respectively, for 4 × super-resolution over the state-of-the-art cubic B-spline. Significantly, our method offers 10x faster inference over the cubic B-spline. The proposed approach for super-resolution of 4D-flow data would potentially improve the subsequent calculation of hemodynamic quantities.

Abstract

Exploiting 4D-flow magnetic resonance imaging (MRI) data to quantify hemodynamics requires an adequate spatio-temporal vector field resolution at a low noise level. To address this challenge, we provide a learned solution to super-resolve in vivo 4D-flow MRI data at a post-processing level. We propose a deep convolutional neural network (CNN) that learns the inter-scale relationship of the velocity vector map and leverages an efficient residual learning scheme to make it computationally feasible. A novel, direction-sensitive, and robust loss function is crucial to learning vector-field data. We present a detailed comparative study between the proposed super-resolution and the conventional cubic B-spline based vector-field super-resolution. Our method improves the peak-velocity to noise ratio of the flow field by 10 and 30% for in vivo cardiovascular and cerebrovascular data, respectively, for 4 × super-resolution over the state-of-the-art cubic B-spline. Significantly, our method offers 10x faster inference over the cubic B-spline. The proposed approach for super-resolution of 4D-flow data would potentially improve the subsequent calculation of hemodynamic quantities.

Statistics

Citations

Dimensions.ai Metrics
7 citations in Web of Science®
6 citations in Scopus®
Google Scholar™

Altmetrics

Downloads

19 downloads since deposited on 20 Feb 2023
15 downloads since 12 months
Detailed statistics

Additional indexing

Item Type:Journal Article, refereed, original work
Communities & Collections:07 Faculty of Science > Department of Quantitative Biomedicine
Dewey Decimal Classification:610 Medicine & health
Scopus Subject Areas:Physical Sciences > Artificial Intelligence
Uncontrolled Keywords:Artificial Intelligence
Language:English
Date:12 August 2022
Deposited On:20 Feb 2023 16:39
Last Modified:28 Apr 2024 01:50
Publisher:Frontiers Research Foundation
ISSN:2624-8212
OA Status:Gold
Free access at:PubMed ID. An embargo period may apply.
Publisher DOI:https://doi.org/10.3389/frai.2022.928181
PubMed ID:36034591
Project Information:
  • : FunderH2020
  • : Grant ID765140
  • : Project TitleJOLT - Harnessing Data and Technology for Journalism
  • Content: Published Version
  • Language: English
  • Licence: Creative Commons: Attribution 4.0 International (CC BY 4.0)