Support for edits #591
Support for edits #591
Conversation
|
Hi, thanks for the pr. It would be great to fix the spelling errors and also add a readme for edits, which can be pretty short in the first iteration, with a 1-2 sentence description of each possible edit. |
|
Spelling should be fixed, documentation will need to be a future thing for now. |
dkuegler
force-pushed
the
feature/edits
branch
from
ad7ed78 to
b24182b
Compare
|
This branch is currently a mess of fixes and features for 6 different things. Those need to be separated into individual branches and PRs. At a minimum, this version seems to work for longitudinal processing (still unclear for edits). |
dkuegler
force-pushed
the
feature/edits
branch
2 times, most recently
from
21f9e19 to
82cd0aa
Compare
m-reuter
force-pushed
the
feature/edits
branch
from
82cd0aa to
8bccee7
Compare
|
I rebased this to current dev which disentangles the edits from the other changes. Please check :
Also update your local with a hard reset. |
dkuegler
force-pushed
the
feature/edits
branch
3 times, most recently
from
69efedc to
0000b25
Compare
dkuegler
force-pushed
the
feature/edits
branch
from
0000b25 to
2bcedbf
Compare
dkuegler
force-pushed
the
feature/edits
branch
7 times, most recently
from
0716df6 to
9aace28
Compare
dkuegler
force-pushed
the
feature/edits
branch
from
9aace28 to
e64fcd0
Compare
|
Maybe there should be a check if the fastsurfer versions are the same between the original run and the rerun? @m-reuter If this is just a warning or a message this will likely get ignored. |
I think warning message in the logs is fine. It should only stop if versions are not compatible any longer (and it would fail somewhere during processing). Otherwise, if someone edits things, that is changing more than most of our versions. |
| 3. Brain extraction by editing `<subject_dir>/mri/brain.mgz` (delta to `brain.auto.mgz`) -> [asegdkt_segfile](#asegdkt_segfile) | ||
| 4. Subcortical segmentation: `<subject_dir>/mri/aseg.presurf.mgz` (delta to `aseg.presurf.auto.mgz`) -> [asegdkt_segfile](#asegdkt_segfile) | ||
| 5. CRS seeds: PONS, CC, LH, RH, watershed => seed files (`<subject_dir>/scripts/seed-(pons|cc|lh|rh|ws).crs.man.dat`) -> [asegdkt_segfile](#asegdkt_segfile) | ||
| 6. WM control points: `<subject_dir>/tmp/control.dat` -> [asegdkt_segfile](#asegdkt_segfile) |
There was a problem hiding this comment.
hm, we could allow them to create control points which should be used by recon-surf when crating wm.mgz?
There was a problem hiding this comment.
I think we could, but there are some comments somewhere why we would not want to do that. I think I did not really understand those comment but just assumed you had decided against this on principle.
There was a problem hiding this comment.
Also, I am not really sure how these labels would propagate back into aseg.mgz and aparc.DKTatlas+aseg.mgz
There was a problem hiding this comment.
They would not, they are only used to increase intensity in thin WM regions to avoid loosing a gyrus or parts of a lobe (usually temporal lobe) due to low grey white contrast there. Let's discuss.
dkuegler
force-pushed
the
feature/edits
branch
from
32174f7 to
bb314af
Compare
- Adds the --edits flag - Add additional log messages for edits - Support for manedit files for asegdkt_segfile. - Also propagate asegdkt_segfile into aseg_segfile and mask_name recon_surf/recon-surf.sh - Adds the --edits flag - Add additional log messages for edits - Change check for existing run into a message instead a fail criteria in edits mode - Do not delete logfile in edits mode - Make mask and aseg_noCC required for edits mode (otherwise it might get unclear what should or was already done between run_fastsurfer and recon-surf) - Add manedit "search" for mask and asegdkt_segfile. recon_surf/talairach-reg.sh - Add an extra argument (new argument #3 to indicate whether talairach.reg should support edits) - Support for edits: some messages and if in edit mode, only run talairach registration if mri/transforms/talairach.xfm and mri/transforms/talairach.auto.xfm are the same (so there are no edits) -- thus edits in talairach.xfm superceed automatic results - talairch-reg.sh will now fail if mri/transforms/talairach.xfm exists, but the script is not running in edits mode (independent if talairach.auto.xfm is the same or not). recon_surf/functions.sh - add the add_file_suffix function, which adds a file suffix (for example manedits) to a file name
…are now possible.
Add what NOT to do section
Co-authored-by: David Kügler <[email protected]>
dkuegler
force-pushed
the
feature/edits
branch
from
bb314af to
01ccce3
Compare
This PR adds support for edits in FastSurfer.
Edits are enabled by the
--editsflag forrun_fastsurfer.sh.Edits supported:
<asegdkt_segfile> via <asegdkt_segfile-w/o-ext>.manedit.
white matter segmentation (wm.mgz) via wm.manedit.mgz
white matter control points (via the standard FreeSurfer control points file)
talairach registration via mri/mri/transforms/talairach.xfm (overwriting talairach registration via mri/mri/transforms/talairach.xfm)
brain.finalsurfs.mgz via brain.finalsurfs.manedit.mgz
This PR also move the allow_root check into functions.sh and introduces run_it and run_it_cmdf function variants that are more clear wrt. argument splitting (see SC2206, SC2207 and similar).