ITK Pre-Processing Algorithm Wrappers in Python/AnisotropicDiffusionVesselEnhancementFilter

Anisotropic Diffusion Vessel Enhancement Filter

This filter is implemented using itkAnisotropicDiffusionVesselEnhancementFilter class. Note that this filter is currently not a part of the ITK source, and the code was obtained from a contribution to the [Insight Journal] by Enquobahrie et al. [1]. The algorithm uses a Hessian-based multi-scale vesselness measure to drive the anisotropic diffusion to happen only along the vessels. This algorithm is slightly slow, and this is one drawback of the algorithm. Currently this algorithm only works for 3D images.

Parameters:

• 1. SigmaMin - this parameter chooses the minimum scale parameter that should be used in the process of finding the vesselness measure. By fixing the value of SigmaMin, effectively all the object of size lesser than SigmaMin are diffused/smoothed out.
• 2. Sigmamax - this parameter chooses the maximum scale parameter used in the vesselness measure. The range [SigmaMin, SigmaMax] should cover the size of the objects (vessels) that we are interested in enhancing.
• 3. NumberOfSigmaSteps - this chooses the number of points in the range [SigmaMin, SigmaMax] that will be used to find the vesselness measure. The points are chosen in a logarithmic fashion.
• 4. NumberOfIterations - this describes the number iterations that should be performed in the anisotropic diffusion process. More the number of iterations, more the smoothing and more diffused the image is.

References

[1] Enquobahrie A., Ibanez L., Bullitt E., and Aylward S., Vessel Enhancing Diffusion Filter, The Insight Journal - 2007 MICCAI Open Science Workshop.

[2] Manniesing R., Viergever M.A., and Niessen W.J., Vessel enhancing diffusion: A scale space representation of vessel structures, Medical Image Analysis, Volume 10, Issue 6, December 2006.