Utilizing Physics-informed neural networks (PINN) to compute 3D blood flow velocity fields from sparse iFR/FFR pressure data sampled across stenosed coronary arteries.
Attaching the codes used for the paper:
Machine Learning Enhanced Hemodynamics: Constructing 3D Blood Flow Fields of Stenosed Coronary Arteries from Pressure Measurements, Siva Viknesh, Ethan Shoemaker, and Amirhossein Arzani
PyTorch codes are included (along with their geometries) for the different examples presented in the paper:
- Idealized Stenosed Coronary Arteries
- Symmetric Stenosis
- Asymmetric Stenosis
- Patient-specific LAD Stenosed Coronary Artery
- Steady Flow
- Transient Flow
Converting the results to VTK: The torch-to-vtk conversion Python programs can be found in PINN- Post Processing folder for both PINN and BC-PINN methodologies.
Installation:
Install Pytorch:
https://pytorch.org/
Install VTK after Pytorch is installed.
An example with pip:
conda activate pytorch
pip install vtk
