GLCM at different distances

Hello @haesleinhuepf !

I come back to you after the great work you did on the implementation of GLCM calculation using CLIJ. The function generateGreyValueCooccurrenceMatrixBox works perfectly and my users and I are really happy of the texture we are able to get !
A recurrent feature is the possibility to measure the GLCM at different distance : the standard is to look at direct neighborhood, but it is also interesting to calculate the same thing at +2, +3, +N pixels away from the central pixel.
As I’m writing, I realize it would have been easier for you to be aware of that parameter earlier, and I’m sorry for that. Would you have time to take a look and implement this modification ?

Thank you very much !

Julien

1 Like

Hey @jdumont ,

I’m glad to hear you like it.

Just to clarify: Is this radius meant to quantify co-occurencene of grey values in all pixels with distance below a given radius or pixels which are approximately at this radius? I’m not so familiar with these operations.

I’m a bit busy in the coming weeks, but added your idea to the wishlist. You don’t have a Java developer who could help out a bit in your vicinity by chance?

Thanks for the feedback :slight_smile:

Cheers,
Robert

1 Like

It would be to quantify pixels which are at this radius.

I completely understand ! I still have my code before CLIJ that I can easily change, it would be slower on the execution but at least it would work, so there is no particular hurry, a place on the wishlist is perfect.

I do, I will check with him if he can have a look. If we end up having something working, I will keep you in touch !

Thank you and have a good day :slight_smile:

Julien

Awesome! Here is the code as a starting point:

Let me know what you think!

Cheers,
Robert

1 Like