Skip to content

Datasets

Giles Tetteh edited this page Apr 10, 2023 · 13 revisions

Welcome to the DeepVesselNet Datasets Page!

We experimented on three main datasets using DeepVesselNet

Synthetic Dataset

This dataset is generated and used for transfer learning. This includes 136 examples of size (325 × 304 × 600). Each example has the following volumes

  1. raw volume which represents the raw intensities. download here
  2. segmentation volume which is the binary segmentation (ground truth) of the vessels (one representing vessels). download here
  3. centerline volume which is a binary segmentation (ground truth) of the centerlines (one representing centerlines). download here
  4. points volume which is a volume containing points. the value representing the number of edges per each point. points with values 3 and above are considered as splitting points. download here
  5. bifurcation volume which is a mask of blocks around the points in the points volume above with values ( > 2) representing bifurcations points. download here
  6. radius volume which gives the radius of the vessels along the centerlines. download here
  7. graph vtk files which contains the vtk graph files used in generating the volumetric images. download here

Time-of-flight (TOF) Magnetic Resonance Angiography (MRA) Data.

This is a real clinical data consisting of 20 training and 20 test sets. Further description and download links will be provided very soon!

Synchrotron Radiation X-ray Tomographic Microscopy

This data is made up of CTA scans of mice brains. Further description and download links will be provided very soon!

For a detailed description of the datasets kindly read the paper below and also cite the same paper in case you use any of the resources in this repository.

@ARTICLE{10.3389/fnins.2020.592352,
  
AUTHOR={Tetteh, Giles and Efremov, Velizar and Forkert, Nils D. and Schneider, Matthias and Kirschke, Jan and Weber, Bruno and Zimmer, Claus and Piraud, Marie and Menze, Björn H.},   
	 
TITLE={DeepVesselNet: Vessel Segmentation, Centerline Prediction, and Bifurcation Detection in 3-D Angiographic Volumes},      
	
JOURNAL={Frontiers in Neuroscience},      
	
VOLUME={14},      
	
YEAR={2020},      
	  
URL={https://www.frontiersin.org/article/10.3389/fnins.2020.592352},       
	
DOI={10.3389/fnins.2020.592352},      
	
ISSN={1662-453X}
}