The University of Auckland
Browse
1/1
3 files

WSSNet: aortic 4D Flow MRI wall shear stress estimation neural network

Version 2 2022-02-07, 23:08
Version 1 2022-02-01, 21:56
software
posted on 2022-02-07, 23:08 authored by Edward FerdianEdward Ferdian
We developed WSSNet, a deep learning method to estimate aortic wall shear stress based on geometry and velocity from 4D Flow MRI.
The network was trained on synthetic data, generated from CFD simulations.

The input and output of the network consists of flattened representation of the aortic vessel surface, translated into velocity sheets and coordinate flatmaps.

Training dataset and pre-trained weights are provided here.
The network was trained using Tensorflow 2.3 with Keras backend.

Files:
1, wssnet: contains the network weights and optimisation parameters to restore/continue training
2. wssnet_dataset: contains the dataset files used for training the network

Please follow the instruction in our github page to use the files.

GitHub:
https://github.com/EdwardFerdian/WSSNet

History

Publisher

University of Auckland

Temporal coverage: start

2022-02-01

Temporal coverage: end

2025-06-26

Usage metrics

    FMHS

    Licence

    Exports

    RefWorks
    BibTeX
    Ref. manager
    Endnote
    DataCite
    NLM
    DC