AutoAWT is a general-purpose wall thickness measurement algorithm for 3D atrial wall thickness measurements, which can be used using binary mask images generated by any segmentation software. This algorithm includes the detection of endocardial and epicardial and wall thickness measurement by the Laplace equation and Euler's method.
The code was written by Oh-Seok Kwon with support from Hui-Nam Pak.
-
Dependencies
- CUDA 10.0 (download link).
-
Datasets repository
- The sample datasets (2 CT and segmented 320 atrial wall binary mask images) for AutoAWT in datasets/Phantom.
- You can build & run (Visual studio 2015).
Step 1. select directory of mask images (datasets/Phantom/masks_1).
Step 2. select directory of DCM (datasets/Phantom/ct)
- Derived results (in "Results" directory)
- WT-endo.plt : Calculated wall thickness from endocardium to epicardium
- WT(projected)-PatientID.plt : Wall thickness projected to surface mesh
- WT(projected)-PatientID.stl : surface mesh
- surface_mesh.stl : Reconstructed surface mesh
- Epi-Endo directory : Images labeled as endocardium, epicardium, and myocardium in binary images
(pixel value 1: endocardium, 2: myocardium, 3: epicardium)
If you find this code useful, please consider citing our paper.
@ARTICLE{9736959,
author={Kwon, Oh-Seok and Lee, Jisu and Park, Je-Wook and Yang, So-Hyun and Hwang, Inseok and Yu, Hee Tae and Shin, Hangsik and Pak, Hui-Nam},
journal={IEEE Access},
title={Three-Dimensional Atrial Wall Thickness Measurement Algorithm From a Segmented Atrial Wall Using a Partial Differential Equation},
year={2022},
volume={10},
number={},
pages={32161-32170},
doi={10.1109/ACCESS.2022.3159795}}
}
If you have any questions or issues using this code, please make a log to the Issue tab.