3
$\begingroup$

I have done fews researches and found out approach such as

  • multiscale and
  • region growing,

that could be used of extracting forearm veins.

However, I tried to use frangi filter and Figure 1 is the segmentation result using the best parameter set I could found;

enter image description here Figure 1

Could anyone suggest what approaches as well the technique to adopt while considering this problem?

I give extra image for further analysis:

enter image description here

$\endgroup$
  • $\begingroup$ well, you listed a lot of approaches, and you even said "my parametrization might be wrong". Why not adjust parameters then? why not try one of these other approaches if Frangi doesn't help (I don't know Frangi, or much about segmentation, but my point is that you're underdefining your question)? $\endgroup$ – Marcus Müller Aug 26 '16 at 9:04
  • 1
    $\begingroup$ i had adjusted the parameter and attached is the best veins extraction that i could obtained, so far i am trying other approaches but i might doing the wrong things and i have only a short time to do it, therefore, i am asking idea from expert here. Thanks again for the comment $\endgroup$ – Anthony Goh Aug 26 '16 at 9:37
  • $\begingroup$ ah, that's a good remark; you could even add that to your question ("this is the segmentation using the best parameter set I could find")! $\endgroup$ – Marcus Müller Aug 26 '16 at 9:38
7
$\begingroup$

So one good step to enhance the vein-like structures is coherence enhancing diffusion:

Weickert, Joachim. "Coherence-enhancing diffusion filtering." International Journal of Computer Vision 31.2-3 (1999): 111-127.

So I first apply this algorithm to your image, aggressively. The next step is to identify the curvilinear structures, which would in this case extract the veins. I already explained how to do this robustly in a previous post. So please check this.

The result is the following: enter image description here

One could apply all the desired noise cleaning to this image. Note that the edges appear very weak, so it is better to tune the parameters, probably even more than I did.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.