Synthetic Dataset Generation, Physical Consistency and Active Learning for Shape-Informed Surrogate Modeling of Hemodynamics
Abstract
Model order reduction has the potential to enable real-time patient-specific assessment of cardiovascular flows. However, it still faces significant challenges stemming from the geometric variability of vascular anatomies. Building upon the SynthAorta dataset (Bosnjak et al., IEEE Trans. Med. Imag. 45(1), 2026) of physiological aortic geometries, we generate an extensive dataset of physiologically consistent computational fluid dynamics solutions through an automatic boundary-condition adjustment procedure. A common reference geometry allows all patient-specific anatomies to be represented by displacement fields, providing point-to-point correspondence without the need for geometric registration. Based on this dataset, we develop Graph U-Net and Transformer U-Net surrogate models to predict velocity and pressure fields at peak systole. We further investigate two strategies to enhance data efficiency and prediction accuracy: active learning and differentiable physical consistency constraints based on the finite element residual. The results show that active learning is strongly regime-dependent: while it offers little benefit over random sampling during early training, it improves performance during fine-tuning through gradient- and subspace-based acquisition strategies. Similarly, physical consistency constraints prove to be substantially more effective as a fine-tuning objective than as a regularizer applied throughout training. Finally, out-of-distribution studies demonstrate that the primary limitation of surrogate models lies in the geometric support of the underlying parametrization rather than in the neural network architectures themselves. The proposed framework combines synthetic dataset generation, geometry-aware surrogate modeling, physics-informed optimization, and active learning into a scalable workflow for inter-patient hemodynamic prediction.
Related
- Machine learning-based thermoelectric materials screening with physical consistency and uncertainty propagation
- Dual Branch Transformer with Physical Priors for PSF Acquisition to Enable PSF Guided Image Restoration against Co Phasing Errors in Optical Synthetic Aperture Interferometric Imaging
- ASTROLABE: A 3D Particle-Level Surrogate for Cosmology Simulation with Physical Flow Matching
- A Two-Dimensional Physics-Informed Neural Network Model of a Hall Effect Thruster Plasma: Self-Consistent Electrostatics and Physical-Consistency Diagnostics
- Variational Quantum Transformer Architecture for Synthetic Language Generation
BibTeX
@misc{romor2026synthetic,
title = {{Synthetic Dataset Generation, Physical Consistency and Active Learning for Shape-Informed Surrogate Modeling of Hemodynamics}},
author = {Francesco Romor and Richard Schussnig and Charles A. Taylor and Gerhard A. Holzapfel and Alfonso Caiazzo},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.27130},
eprint = {2610.27130},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.27130},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Synthetic Dataset Generation, Physical Consistency and Active Learning for Shape-Informed Surrogate Modeling of Hemodynamics
AU - Francesco Romor
AU - Richard Schussnig
AU - Charles A. Taylor
AU - Gerhard A. Holzapfel
AU - Alfonso Caiazzo
PY - 2026
DA - 2026/10/19
PB - arXiv
DO - 10.48550/arxiv.2610.27130
UR - https://arxiv.org/abs/2610.27130
AB - Model order reduction has the potential to enable real-time patient-specific assessment of cardiovascular flows. However, it still faces significant challenges stemming from the geometric variability of vascular anatomies. Building upon the SynthAorta dataset (Bosnjak et al., IEEE Trans. Med. Imag. 45(1), 2026) of physiological aortic geometries, we generate an extensive dataset of physiologically consistent computational fluid dynamics solutions through an automatic boundary-condition adjustment procedure. A common reference geometry allows all patient-specific anatomies to be represented by displacement fields, providing point-to-point correspondence without the need for geometric registration. Based on this dataset, we develop Graph U-Net and Transformer U-Net surrogate models to predict velocity and pressure fields at peak systole. We further investigate two strategies to enhance data efficiency and prediction accuracy: active learning and differentiable physical consistency constraints based on the finite element residual. The results show that active learning is strongly regime-dependent: while it offers little benefit over random sampling during early training, it improves performance during fine-tuning through gradient- and subspace-based acquisition strategies. Similarly, physical consistency constraints prove to be substantially more effective as a fine-tuning objective than as a regularizer applied throughout training. Finally, out-of-distribution studies demonstrate that the primary limitation of surrogate models lies in the geometric support of the underlying parametrization rather than in the neural network architectures themselves. The proposed framework combines synthetic dataset generation, geometry-aware surrogate modeling, physics-informed optimization, and active learning into a scalable workflow for inter-patient hemodynamic prediction.
ER -