Scalable Neuroimaging Processing Engine for Minipig MRI

Introduction

PigSNIPE is a software package for analysis of Minipig brain Magnetic Resonance Images (MRI). It is developed by the SINAPSE lab at the University of Iowa, Electrical and Computer Engineering department.

PigSNIPE provides a fully automatic pipeline that allows for image registration, AC-PC alignment, brain mask segmentation, skull stripping, tissue segmentation, caudate-putamen brain segmentation, and landmark detection.

For detailed information we refer you to the PigSNIPE paper.

Setup

1. Clone this git repository.

   $ git clone https://github.com/BRAINSia/PigSNIPE.git

2. Create a virtual environment and install required packages.

   $ python3 -m venv <path_to_virtual_env>

   $ source <path_to_virtual_env>/bin/activate

   $ cd <path_to_repo>

   $ pip install -r requirements.txt

3. Build BRAINSTools

   Refer to the BRAINSTools GitHub repository for specific build instructions ("Building" section).

   NOTES:

   • to optimize the build for your machine hardware, use 'NATIVE' mode in ccmake setup BRAINSToools_CXX_OPTIMIZATION_ -mtune=native -march=native BRAINSToools_C_OPTIMIZATION_FL -mtune=native -march=native
   • We advise to build BRAINSTools as default. However it is a large package. For lightweight build, set the following to 'ON' and the rest to 'OFF':
     • USE_BRAINSFit
     • USE_BRAINSResample
     • USE_BRAINSConstellationDetector

4. Setting Up Binaries and Libraries.

   To set up binaries and libraries needed to utilize this repository, run the following script inside the PigSNIPE repository.

   $ python3 setup.py -b <path_to_BRAINSTools_build_dir>

5. Download the zip file containing model weights.

   Note: The Link Will be available shortly. If you wish to use the tool sooner email [email protected] or [email protected].

6. Unzip the DL_MODEL_PARAMS.zip directory in the cloned repo.

   $ unzip <path_to_zip_file> -d <path_to_repo>

7. Run PigSNIPE pipeline

   $ python3 pigsnipe -> You should the help message for the script.

   The example command to run the pipeline is:

   $ python3 pigsnipe -t1 <path_to_T1w> -t2 <path_to_T2w> -o <path_to_result_directory> --keep_temp_files

Authors

Michal Brzus - Ph.D. student at the University of Iowa

Hans Johnson, Ph.D - Professor at the University of Iowa, Electrical and Computer Engineering department.

for contact, email: [email protected]