Problems with SNOW algoritm

[alabaien]

Dear network,

I am experiencing some difficulties while implementing the SNOW algorithm in one of my binary images. As you could see, it oversegmentates (from my point of view of course) the pore space the region, especially in the big elliptical pore colored by green and yellow

Would you consider a good segmentation of the pore space? Would consider touching other variables such blurring or the radius R for searching the maxima points in the Euclidean Distance map?

[alabaien]

Thanks in advance :)

[jgostick]

That looks fine to me. those 2 pores will basically act as one. Flow will move between them as if they are a single long pore.

[alabaien]

OK! So it won´t give any problem with flow simulations.Thanks!

[jgostick]

Exactly. In fact, two pores in there is better, because the alternative is one pore in the middle, connected to the neighboring pores by long skinny throats, which will create resistance to flow.