External Email - Use Caution
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 pial file.
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 ( https://secure-web.cisco.com/1NYZrvHyVXqxTpZBiwiJ49J4EeeNWwqM6oBe0m-ipN8GosM...) 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 ( https://secure-web.cisco.com/1eQ5Q62ZQtG-RYtwAFFKvhYNEYp7A20Vte8M_v6fEXDVssP...). This image is of surf/lh.pial with 149,968 vertices and 299,932 faces. [image: 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. ( https://secure-web.cisco.com/1JB4zyZhlTjmxoItAOFNb40IoMYhEZt8T-O0MINoKqg8clI...). This image is from the same T1 image as above, but processed with FS7, resulting in 147,292 vertices and 294,580 faces. [image: rh_pial_oblique_from_fs7.png]
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
