![[2408.07110] Physics-informed graph neural networks for flow field estimation in carotid arteries](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2408.07110] Physics-informed graph neural networks for flow field estimation in carotid arteries
Summary
This article presents a novel approach using physics-informed graph neural networks to estimate hemodynamic flow fields in carotid arteries, leveraging limited data from 4D flow MRI.
Why It Matters
The research addresses the challenge of non-invasive measurement of hemodynamic quantities, crucial for understanding cardiovascular diseases. By integrating machine learning with physics principles, the study enhances the accuracy of flow estimations, potentially improving patient outcomes in cardiovascular health.
Key Takeaways
- Introduces a surrogate model for estimating blood flow fields using graph neural networks.
- Reduces the need for large datasets by utilizing physics-informed priors.
- Demonstrates effective transfer of learned models to different vascular geometries.
- Combines PointNet++ architecture with group-steerable layers for improved performance.
- Validates the model's accuracy through extensive experiments in carotid arteries.
Quantitative Biology > Quantitative Methods arXiv:2408.07110 (q-bio) [Submitted on 13 Aug 2024 (v1), last revised 20 Feb 2026 (this version, v2)] Title:Physics-informed graph neural networks for flow field estimation in carotid arteries Authors:Julian Suk, Dieuwertje Alblas, Barbara A. Hutten, Albert Wiegman, Christoph Brune, Pim van Ooij, Jelmer M. Wolterink View a PDF of the paper titled Physics-informed graph neural networks for flow field estimation in carotid arteries, by Julian Suk and 5 other authors View PDF HTML (experimental) Abstract:Hemodynamic quantities are valuable biomedical risk factors for cardiovascular pathology such as atherosclerosis. Non-invasive, in-vivo measurement of these quantities can only be performed using a select number of modalities that are not widely available, such as 4D flow magnetic resonance imaging (MRI). In this work, we create a surrogate model for hemodynamic flow field estimation, powered by machine learning. We train graph neural networks that include priors about the underlying symmetries and physics, limiting the amount of data required for training. This allows us to train the model using moderately-sized, in-vivo 4D flow MRI datasets, instead of large in-silico datasets obtained by computational fluid dynamics (CFD), as is the current standard. We create an efficient, equivariant neural network by combining the popular PointNet++ architecture with group-steerable layers. To incorporate the physics-informed priors, we derive...
