An extension module in 3DSlicer for COVID19 CT segmentation using Convolutional Neural Networks.
We trained a 2D ConvNet to segment lung and possible infected regions from patient CT scans. We then incorporated this ConvNet segmentation model into 3DSlicer.
This module requires Keras/Tensorflow, scikit-image, scipy, and numpy (all installed under 3DSlicer).
Location of the model can be changed in COVID19_CT_SEGMENTATION/ConvNetCovid/ConvNetCovid.py:
line 55: segmentation_model = load_model('xxx_Model.h5')
