A comparison of machine learning methods for recovering noisy and missing 4D flow MRI data
International Journal for Numerical Methods in Biomedical Engineering · 2024
DOI: 10.1002/cnm.3858 ↗Kurz gefasst
Vergleichsstudie in der Zeitschrift International Journal for Numerical Methods in Biomedical Engineering, 2024. Verfasst von Hunor Csala und 3 weiteren Autor:innen. Beteiligte Einrichtung: Department of Mechanical Engineering, University of Utah, Salt Lake City, Utah, USA. Hauptthemen laut MEDLINE: Machine Learning, Magnetic Resonance Imaging, Algorithms. Die Arbeit wurde bisher 5-mal zitiert (Crossref) und nennt 101 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC 4.0).
Abstract (Originalsprache)
Experimental blood flow measurement techniques are invaluable for a better understanding of cardiovascular disease formation, progression, and treatment. One of the emerging methods is time‐resolved three‐dimensional phase‐contrast magnetic resonance imaging (4D flow MRI), which enables noninvasive time‐dependent velocity measurements within large vessels. However, several limitations hinder the usability of 4D flow MRI and other experimental methods for quantitative hemodynamics analysis. These mainly include measurement noise, corrupt or missing data, low spatiotemporal resolution, and other artifacts. Traditional filtering is routinely applied for denoising experimental blood flow data without any detailed discussion on why it is preferred over other methods. In this study, filtering is compared to different singular value decomposition (SVD)‐based machine learning and autoencoder‐type deep learning methods for denoising and filling in missing data (imputation). An artificially corrupted and voxelized computational fluid dynamics (CFD) simulation as well as in vitro 4D flow MRI data are used to test the methods. SVD‐based algorithms achieve excellent results for the idealized case but severely struggle when applied to in vitro data. The autoencoders are shown to be versatile and applicable to all investigated cases. For denoising, the in vitro 4D flow MRI data, the denoising autoencoder (DAE), and the Noise2Noise (N2N) autoencoder produced better reconstructions than filtering both qualitatively and quantitatively. Deep learning methods such as N2N can result in noise‐free velocity fields even though they did not use clean data during training. This work presents one of the first comprehensive assessments and comparisons of various classical and modern machine‐learning methods for enhancing corrupt cardiovascular flow data in diseased arteries for both synthetic and experimental test cases.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Department of Mechanical Engineering, University of Utah, Salt Lake City, Utah, USA.
- Scientific Computing and Imaging Institute, University of Utah, Salt Lake City, Utah, USA.
- Department of Mechanical, Industrial and Manufacturing Engineering, University of Toledo, Toledo, Ohio, USA.
- Department of Mechanical Engineering, University of Wisconsin-Milwaukee, Milwaukee, Wisconsin, USA.
ORCID-Profile: Hunor Csala, Amirhossein Arzani
Zitieren
APA
Csala, H., Amili, O., D'Souza, R. M., & Arzani, A. (2024). A comparison of machine learning methods for recovering noisy and missing 4D flow MRI data. International Journal for Numerical Methods in Biomedical Engineering. https://doi.org/10.1002/cnm.3858
Vancouver
Csala H, Amili O, D'Souza RM, Arzani A. A comparison of machine learning methods for recovering noisy and missing 4D flow MRI data. International Journal for Numerical Methods in Biomedical Engineering. 2024. doi:10.1002/cnm.3858
BibTeX
@article{Csala2024,
author = {Csala, Hunor and Amili, Omid and D'Souza, Roshan M and Arzani, Amirhossein},
title = {A comparison of machine learning methods for recovering noisy and missing 4D flow MRI data},
journal = {International Journal for Numerical Methods in Biomedical Engineering},
year = {2024},
doi = {10.1002/cnm.3858}
}
EUVIMED