Caret-users,
A short while back, Roland Marcus Rutschmann posted about partial hem
flattening trouble. It took a while to resolve the issues, and most
messages traded were (mercifully) off-list.
Here is a recap for the benefit of future troubleshooters entering
keywords into search engines:
* Caret was extending the last slice, instead of second to last slice;
John has fixed this buglet. For details, see:
http://brainmap.wustl.edu/pub/donna/ARCHIVE/BAD_LAST_SLICE/bad_last_slice.html
login pub
password download
* Caret was using the raw surface for painting CUT.FACE (not all of the
padding was painted red); John has fixed this buglet.
* The initial flat map still looks weird, but this isn't a problem. See
DVE's message below.
Donna
-------- Original Message --------
Subject: Re: Initial flat map issues
Date: Sat, 1 Dec 2007 15:11:40 -0600 (CST)
From: [EMAIL PROTECTED]
To: David Van Essen <[EMAIL PROTECTED]>
CC: Roland Marcus Rutschmann
<[EMAIL PROTECTED]>, John Harwell
<[EMAIL PROTECTED]>, Donna Dierker <[EMAIL PROTECTED]>,
David Van Essen <[EMAIL PROTECTED]>
Just one inline reply near the end of DVE's reply.
Roland et al.,
A few additional comments.
1) I agree that you shouldn't need to remove the big marginal tiles
in order to get a clean flat map.
2) As Donna noted, the marginal tiles should now be a problem ONLY if
the 'north pole' that is used as the projection-point for the initial
flattening lies over a non-red (non-cut-face node).
If that were to occur in a different case, the quick fix would be to
rotate the sphere slightly, apply the current view, and then proceed
with flattening.
If it were a common occurrence, then the problem can be fixed more
generically by padding for additional slices, so that a higher
fraction of the final spherical surface is occupied by red. However,
I don't think this is worth doing until we know that the need is real.
3) Regarding whether to cut or not, I agree that the uncut map looks
'nicer'. On the other hand, it definitely has greater compression of
the occipital pole, so you have to decide where your tradeoffs lie.
4) In this regard, it would be helpful to know what types of analyses
your are doing and how you are comparing results across individuals.
Is there a fundamental reason why you are flattening only the
occipital lobe instead of full-hemisphere segmentation?
I am particularly curious as to whether you envision registering your
individual hemisphere datasets to our PALS atlas. This can be done
at present, but the process is tedious. I have an outline of a plan
to make partial hemisphere registration faster, easier and more
reliable, but it hasn't gotten onto mine or John's front burner. Is
this something that would be of interest to you?
Meanwhile, here is a script that worked for a monkey occipital lobe:
http://brainvis.wustl.edu/pub/donna/RISHI/PARTIAL_HEM_REG/
login pub
password download
David
On Nov 30, 2007, at 1:57 PM, Donna Dierker wrote:
Good point: If you make no cuts (and I agree you don't need to --
either the calcarine or north pole rectangulars), then Caret will
name the topo for you.
On 11/30/2007 01:55 PM, John Harwell wrote:
On Nov 30, 2007, at 12:39 PM, Roland Marcus Rutschmann wrote:
On Thursday 29 November 2007 22:58, John Harwell wrote:
Roland,
Hi again,
We have fixed the problem with the padding of the partial
hemisphere's cut face.
just some first results before I leave for the weekend.
I tried your new version on my previous segmentation with and
without cropping
the last (most anterior) slice. I have no crossovers all over the
place.
Actually less crossovers without cropping the last slice. (see
cropped and
uncropped jpegs).
After cutting the large rectangulars away the morphed flat looks
good to me. I
tried with and without applying a cut to the calcarine. The one
without cut
looks better to me. What do you think. Would I save such an
"uncut" flat
surface as "open" or "cut" in the topology file save dialogue?
It may not be necessary to remove the "large rectangulars" as the
multi-resolution morphing may take care of them.
My vote is to save the topology as "CUT" as Caret automatically
assigns cut topology to flat surfaces.
John
