This is a git repo with the implementation in pytorch of the methodology proposed 
in the MICCAI 2019 paper entitled "A Deep Learning Design for improving Topology
Coherence in Blood Vessel Segmentation"

To try the code, I recommend downloading the DRIVE dataset, following the 
instructions given in the following page:
https://www.isi.uu.nl/Research/Databases/DRIVE/

This dataset is already organized as the code is expecting, such that it
suffices to copy it as detailed below:

root - code
     - README.txt
     - LICENSE.txt
     - data - DRIVE - training
		    - test

Nonetheless, it is straightforward to adapt the code to any particular needs.

For any enquiry contact me at one of the following emails:
[email protected]
[email protected]