Long computation time for distance map from Mesh 3D ROI

26 views
Skip to first unread message

Julien Pontabry

unread,
Mar 9, 2017, 10:03:27 AM3/9/17
to Icy imaging
Hi all,

I am doing distance-related quantifications and it is working quite well, but I have a computation time issue with any distance-based plugins (like Distance Map or Inclusion ROI plugins).

The computation time is usually long (more than 15 minutes for one single confocal stack and one channel) when I am using ROI than have been convexified with the plugin Convexify.
It appears that without using the Convexify plugin prior to distances calculation, it is quite fast (at most a few seconds). However, using the convexification does really improve the segmentation.

I saw that the only difference between before and after convexification is that the convexified ROI are in fact 3D Meshes. So I suspect that it is an issue related to the 3D Mesh to 3D mask conversion (for the distance map computation).

Does anyone have any idea on how to speed up this step?

Thanks in advance for your help!

Julien.

Alexandre Dufour

unread,
Mar 24, 2017, 1:53:04 PM3/24/17
to Icy imaging
Dear Julien,
I’ve just published a quick fix for this issue, but didn’t have time to test it. Could you test and report back?
Cheers
Alexandre

Julien Pontabry

unread,
Apr 3, 2017, 5:13:32 AM4/3/17
to Icy imaging
Hi Alexandre,

Thanks for the fix. I tried it on a few images and it seems to be a bit faster. However I need more trials to give you a speed increase.

Bests,
Julien.
Reply all
Reply to author
Forward
0 new messages