Adding L3CNS_Wood2018 template https://virtualflybrain.org/reports/VFB_00049000

[Robbie1977]

Adding L3 template to allow mirroring as a start

[schlegelp]

Hi Robbie. Is this ready to be merged? Just wondering b/c you added a file with mirror landmarks but didn't add a mirror registration?

[schlegelp]

For the landmarks: are the coordinates in microns (i.e. matching the template)? The values seem to go in the tens of thousands which seems unrealistic for um space.

[Robbie1977]

Let me run some checks first.

The template is here if you do have a quick process to calculate the mirror: https://v2a.virtualflybrain.org/data/VFB/i/0004/9000/VFB_00049000/volume.nrrd

[schlegelp]

Is the template already mirror symmetrical?

[Robbie1977]

Almost but isn't perfect

[schlegelp]

Yeah, I can see that. Normally, I would flip the stack along its x-axis and then register flipped stack against the original. However, the stack seems to strongly label tracts, nerves and cell-cell contacts - an anti-Fasciclin staining perhaps? I worry that these features won't work well for a registration.

I also noticed some odd rectangular truncation in the volume:

[Robbie1977]

It's neurotactin, but this usually aligns reasonably with CMTK, so the flip should [fingers crossed] work. It's made from merged tiles, hence the straight borders but should be enough boundary to align to, plus the neurotactin should be symmetrical.