All publications

Thesis / 2023

Estimation of FFR in coronary arteries with deep learning

Patryk Rygiel

Wrocław University of Science and Technology · Published

A synthetic coronary mesh is converted to a point cloud and geometric features for learning hemodynamic predictions from CFD.
The thesis workflow learns hemodynamic fields from coronary surface point clouds, geometric features and synthetic CFD reference data. Figure 4.1 · Rygiel (2023) · Patryk Rygiel · CC BY 4.0

MSc thesis

This thesis investigates estimating virtual fractional flow reserve from coronary artery geometry using deep learning. It develops a point-cloud approach that combines explicit vessel features with learned geometric representations and evaluates its predictions against reference flow simulations.

The thesis was completed in 2023 under the supervision of Maciej Zięba. The public arXiv version was deposited in 2024. The related CenterlinePointNet++ conference paper is listed as a separate publication.

Cite this work

@mastersthesis{ffrthesis2023,
  title = {Estimation of FFR in coronary arteries with deep learning},
  author = { Patryk Rygiel },
  year = {2023},
  school = {Wrocław University of Science and Technology},
  eprint = {2405.12224},
  archivePrefix = {arXiv},
  url = { https://arxiv.org/abs/2405.12224 }
}