Drawing Annotation

109 views
Skip to first unread message

Sid J.

unread,
Jan 23, 2019, 1:52:54 PM1/23/19
to Neuroglancer
Hi, 

I was looking into how to draw more complex annotation (similar to ones displayed in the 3rd FAFB-demo available on GitHub). I was wondering how you would recommend going about it and if there is already an existing functionality for this (I have seen the point annotations to create points, lines, spheres etc.).  

Thanks!

Jeremy Maitin-Shepard

unread,
Jan 23, 2019, 2:28:40 PM1/23/19
to Sid J., Neuroglancer
Unfortunately there is no support for creating arbitrary mesh/polyhedra annotations within neuroglancer, although that could be a useful feature to add if you were interested.  In the FAFB demo, existing triangulated meshes (from https://jefferis.github.io/elmr/reference/FAFB.surf.html) were converted to the neuroglancer precomputed mesh format (for the "neuropil-regions-surface" layer) and to the VTK ASCII mesh format (for the "neuropil-full-surface" layer).  The UI doesn't yet support a way to create mesh layers as shown in the example, but if you look at the state JSON (click on the {} icon at the top right) you can see how to do it.

If you have an existing triangulated mesh, you may find the https://github.com/seung-lab/cloud-volume package useful for converting to the precomputed mesh format.  I've also attached the scripts used to convert those meshes for the FAFB demo.  The source format was a very unusual format (R language variable export format), but they might be useful as examples nonetheless.
convert_surf.py
convert_surf2.py
Message has been deleted

Jeremy Maitin-Shepard

unread,
Jan 25, 2019, 3:09:09 PM1/25/19
to Sid J., Neuroglancer
Neuroglancer doesn't currently support computing and rendering cross-sections of a triangulated mesh in the sliceviews, although it would be possible to support.  Currently in order to see cross sections you would have to also compute a volumetric representation (e.g. in the precomputed volume format).

On Fri, Jan 25, 2019 at 12:03 PM Sid J. <ug201...@iitj.ac.in> wrote:
Thanks a lot the scripts were extremely useful and I was able to import and visualize the mesh. 

I noticed that the projections of the mesh do not show up in the cross-sectional views and on cross-checking with the demos I see that is available for annotations in type segmented format. Am I missing something here? In case this is not available could you give me an idea as to what would be required to add this functionality or would it be easier to convert the mesh into the precomputed format and add it as a segmented mask?



On Wednesday, January 23, 2019 at 11:28:40 AM UTC-8, Jeremy Maitin-Shepard wrote:
Unfortunately there is no support for creating arbitrary mesh/polyhedra annotations within neuroglancer, although that could be a useful feature to add if you were interested.  In the FAFB demo, existing triangulated meshes (from https://jefferis.github.io/elmr/reference/FAFB.surf.html) were converted to the neuroglancer precomputed mesh format (for the "neuropil-regions-surface" layer) and to the VTK ASCII mesh format (for the "neuropil-full-surface" layer).  The UI doesn't yet support a way to create mesh layers as shown in the example, but if you look at the state JSON (click on the {} icon at the top right) you can see how to do it.

If you have an existing triangulated mesh, you may find the https://github.com/seung-lab/cloud-volume package useful for converting to the precomputed mesh format.  I've also attached the scripts used to convert those meshes for the FAFB demo.  The source format was a very unusual format (R language variable export format), but they might be useful as examples nonetheless.

--
You received this message because you are subscribed to the Google Groups "Neuroglancer" group.
To unsubscribe from this group and stop receiving emails from it, send an email to neuroglancer...@googlegroups.com.
To post to this group, send email to neurog...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/neuroglancer/038dec09-e14f-4d6e-92e3-6e9a2650b17b%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
Reply all
Reply to author
Forward
0 new messages