Skip to content

FastSurfer Workflow

Sandip Samal edited this page Sep 24, 2020 · 14 revisions

pl-fshack -> pl-fastsurfer_inference -> pl-mgz2lut_report

********************************* FastSurfer Workflow ******************************************

Step 1) Pull all the docker images

docker pull fnndsc/pl-fshack         \
docker pull fnndsc/pl-fastsurfer_inference \
docker pull fnndsc/pl-mgz2lut_report

Step 2) Get input data for 'pl-fshack' & run

cd ~/                                              
mkdir devel                                       
cd devel                                          
export DEVEL=$(pwd)                               
git clone https://github.com/FNNDSC/SAG-anon-nii.git

docker run --rm                                                         \
-v ${DEVEL}/SAG-anon-nii/:/incoming -v ${DEVEL}/results/:/outgoing  \
fnndsc/pl-fshack fshack.py                                          \
-i SAG-anon.nii                                                     \
-o recon-of-SAG-anon-nii                                            \
--exec recon-all                                                    \
--args 'ARGS: -autorecon1'                                    \
/incoming /outgoing

Step 3) Run 'pl-fastsurfer_inference' on the results of the previous plugin

docker run --rm                                                         \
-v ${DEVEL}/results/recon-of-SAG-anon-nii/mri/:/incoming -v ${DEVEL}/inference_results/:/outgoing  \
fnndsc/pl-fastsurfer_inference fastsurfer_inference.py                                          \
--t .                                                     \
--in_name brainmask.mgz                                   \                    
/incoming /outgoing

Step 4) Run 'pl-mgz2lut_report' on the results of the previous plug-in

docker run --rm                                                         \
-v ${DEVEL}/inference_results/:/incoming -v ${DEVEL}/mgz_reports/:/outgoing  \
fnndsc/pl-mgz2lut_report mgz2lut_report.py                                          \
--file_name aparc.DKTatlas+aseg.deep.mgz                                                     \
--report_name simpleMGZReport                                   \ 
--report_types csv,json,pdf,txt                                    \                
/incoming /outgoing

Run the work-flow on CLI using CUBE

Set some environment variables

Clear pman

Push data to Swift

Run pl-dircopy

Run pl-fastsurfer_inference

Run pl-mgz2lut_report

Clone this wiki locally