External Email - Use Caution        

Thank you for your response, Doug. I am including two screenshots from FreeView side-by-side. These are from the same two lh.pial files mentioned in the first email.
freeview_6-vs-7.png
They are very very similar and render correctly in FreeView. But when I set their transparency from 1.00 to 0.25, similar artifacts appear.
freeview_6-vs-7-25a.png
It's difficult to say which is "right" from this. There's nothing behind the pial surface to exploit the transparency. But there's a distinct difference between the two versions.

Thanks again. Happy to assist in figuring this out in any way I can,
    Mike Schmidt


On Mon, Feb 15, 2021 at 12:10 PM Douglas N. Greve <dgreve@mgh.harvard.edu> wrote:
Do these show up in freeview? You can turn on the mesh and vertices

On 2/14/2021 8:48 PM, Mike Schmidt wrote:

        External Email - Use Caution        

Hello FreeSurfer Developers and Experts,

I have written a script that reads the binary representations of the vertices and faces from FreeSurfer's surf/{l,r}h.pial files and saves them out unchanged to stl and obj file formats for 3D rendering. I have used it with FreeSurfer 6 and several MRI images, everything working fine. Processing the identical input, a single T1-weighted MPRAGE MRI, FreeSurfer 7 works great and finishes faster than FreeSurfer 6, but we see strange aliasing on the 3D objects. The aliasing is identical if I avoid all of my own code and use mris_convert (from either FreeSurfer version) to generate an stl file from the lh.pial and rh.pial files.

I've spent a lot of time debugging, messing with different materials and textures, generating vertex normals at different stages, but I can't find anything that makes any difference except the FreeSurfer version. I have seen several items in the release notes (MailScanner has detected a possible fraud attempt from "secure-web.cisco.com" claiming to be https://surfer.nmr.mgh.harvard.edu/fswiki/ReleaseNotes) that may be related, but I'm not sure how to go about investigating them.

Left cortex images are included, with links if they don't show up in the email, of the same brain processed with FS6 and FS7. These are rendered in Firefox with the three.js library.

FreeSurfer 6.0.0 works well with a smooth surface (MailScanner has detected a possible fraud attempt from "secure-web.cisco.com" claiming to be https://dragonfruit.mfs.ms/rh_pial_oblique_from_fs6.png). This image is of the correctly rendered surf/lh.pial with 149,968 vertices and 299,932 faces.
rh_pial_oblique_from_fs6.png

FreeSurfer 7.1.1 results in the same size and shape cortex, but renders strangely with mottled shading of triangles. (MailScanner has detected a possible fraud attempt from "secure-web.cisco.com" claiming to be https://dragonfruit.mfs.ms/rh_pial_oblique_from_fs7.png). This image is from the same T1 image as above, but processed with FS7.1.1, resulting in 147,292 vertices and 294,580 faces.
rh_pial_oblique_from_fs7.png

I have run this on several T1 files with similar results on each. FreeSurfer 6 surfaces always look smooth and refract light nicely. FreeSurfer 7 surfaces do not. I'm out of ideas, and the only difference I can find is FreeSurfer version, so thank you to anyone who can point me in the right direction to fix this issue.

Mike Schmidt

_______________________________________________
Freesurfer mailing list
Freesurfer@nmr.mgh.harvard.edu
MailScanner has detected a possible fraud attempt from "secure-web.cisco.com" claiming to be https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer

_______________________________________________
Freesurfer mailing list
Freesurfer@nmr.mgh.harvard.edu
MailScanner has detected a possible fraud attempt from "secure-web.cisco.com" claiming to be https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer