LONG: long_fastsurfer.sh¶
Note
Please also see the documentation on Longitudinal Processing.
Usage help text¶
$ ./long_fastsurfer.sh --help
Setting ENV variable FASTSURFER_HOME to script directory /home/runner/work/FastSurfer/FastSurfer/src.
Change via environment to location of your choice if this is undesired (export FASTSURFER_HOME=/dir/to/FastSurfer)
Usage: long_fastsurfer.sh --tid <tid> --t1s <T1_1> <T1_2> .. --tpids <tID1> <tID2> .. [OPTIONS]
long_fastsurfer.sh takes a list of T1 full head image and sequentially creates:
(i) a template directory for the specific person
(ii) directories for each processed time point in template space,
here you find the final longitudinal results
FLAGS:
--tid <templateID> ID for person-specific template directory inside
$SUBJECTS_DIR to be created"
--t1s <T1_1> <T1_2> .. T1 full head inputs for each time point (do not need
to be bias corrected). Requires ABSOLUTE paths!
--tpids <tID1> <tID2> .. IDs for future time points directories inside
$SUBJECTS_DIR to be created later (during --long)
--t2s <T2_1> <T2_2> .. *Optional* T2 full head inputs, one per time point in
the order of --tpids, for the hypothalamus module.
Every time point needs one, or none does. Each is
registered to its time point's T1 in template space,
or with --reg_mode none, taken as co-registered with
the T1 given in --t1s and mapped into template space
with that T1's transform. Only the long_seg stage
uses them. Requires ABSOLUTE paths!
--sd <subjects_dir> Output directory $SUBJECTS_DIR (or pass via env var)
--py <python_cmd> Command for python, used in both pipelines.
Default: "python3 -s"
(-s: do no search for packages in home directory)
-h --help Print Help
Stage control:
--stage <stage> Run specific stage(s). Can be specified multiple times.
Valid stages: prepare, template_seg, template_surf,
long_seg, long_surf, all
Default: all (runs full pipeline)
Stage dependencies:
- prepare: none (requires --t1s and --tpids)
- template_seg: prepare
- template_surf: prepare, template_seg
- long_seg: prepare
- long_surf: prepare, template_seg, template_surf, long_seg
These take the place of --seg_only and --surf_only:
segmentation only: prepare, template_seg, long_seg
surfaces only: template_surf, long_surf
A dependency that already exists on disk does not have
to be listed again, so drop prepare on a re-run.
Parallelization options:
All of the following options will activate parallel processing of the template and the longitudinal time-point images
where possible. Additionally, the number of different processes for segmentation and surface reconstructionis set.
--parallel <n>|max See above, sets the size of the processing pool for segmentation and surface reconstruction
--parallel_seg <n>|max See above, only sets the size of the processing pool for segmentation (default: 1)
--parallel_surf <n>|max See above, only sets the size of the processing pool for surface reconstruction (default: 1)
--t1, --t2 and --sid are not accepted here, they are populated per time point from --t1s, --t2s
and --tpids.
--seg_only and --surf_only are not accepted either, pick the stages above instead. Every other
run_fastsurfer.sh option is supported, see 'run_fastsurfer.sh --help'.
REFERENCES:
If you use this for research publications, please cite:
For FastSurfer (both):
Henschel L, Conjeti S, Estrada S, Diers K, Fischl B, Reuter M, FastSurfer - A
fast and accurate deep learning based neuroimaging pipeline, NeuroImage 219
(2020), 117012. https://doi.org/10.1016/j.neuroimage.2020.117012
Henschel L*, Kuegler D*, Reuter M. (*co-first). FastSurferVINN: Building
Resolution-Independence into Deep Learning Segmentation Methods - A Solution
for HighRes Brain MRI. NeuroImage 251 (2022), 118933.
http://dx.doi.org/10.1016/j.neuroimage.2022.118933
And for longitudinal processing:
Reuter M, Schmansky NJ, Rosas HD, Fischl B. Within-subject template estimation
for unbiased longitudinal image analysis, NeuroImage 61:4 (2012).
https://doi.org/10.1016/j.neuroimage.2012.02.084
For cerebellum sub-segmentation:
Faber J*, Kuegler D*, Bahrami E*, et al. (*co-first). CerebNet: A fast and
reliable deep-learning pipeline for detailed cerebellum sub-segmentation.
NeuroImage 264 (2022), 119703.
https://doi.org/10.1016/j.neuroimage.2022.119703
For hypothalamus sub-segemntation:
Estrada S, Kuegler D, Bahrami E, Xu P, Mousa D, Breteler MMB, Aziz NA, Reuter M.
FastSurfer-HypVINN: Automated sub-segmentation of the hypothalamus and adjacent
structures on high-resolutional brain MRI. Imaging Neuroscience 2023; 1 1–32.
https://doi.org/10.1162/imag_a_00034