Multiparametric Image Analysis

MIA is shorthand for “Multiparametric Image Analysis”. It is intended to be a complete image processing environment mainly targeted at the analysis and visualization of large amounts of MRI data.

MRI data analysis often requires a complex succession of data processing pipelines applied to a set of data acquired in an MRI exam or over several MRI exams. This analysis may need to be repeated a large number of times in studies involving a large number of acquisition sessions. Such that manual execution of the processing modules or simple ad-hoc scripting of the process may become error-prone, cumbersome and difficult to reproduce. Data processing pipelines exist in separate heterogeneous toolboxes, developed in-house or by other researchers in the field. This heterogeneity adds to the complexity of the modules are to be invoked manually.

Populse_MIA aims to provide easy tools to perform complex data processing based on a definition of the inputs and outputs of the individual pipelines on a conceptual level, and implies identifying data with respect to their role in an analysis project: “the scan type”, “the subject being scanned”, “the group of this subject is part of”, etc.

Installation

Usage

  • After an installation in user mode:

    python3 -m populse_mia

  • After an installation in developer mode, interprets the main.py file from the source code directory:

    cd [populse_install_dir]/populse_mia/python/populse_mia  
python3 main.py

  • Depending on the operating system used, it was observed some compatibility issues with PyQt5/SIP. In this case, we recommend, as a first attempt, to do:

    python3 -m pip install --force-reinstall pyqt5==5.14.0
python3 -m pip install --force-reinstall PyQt5-sip==5.0.1

Contributing to the project

If you'd like to contribute to the project please read our developer documentation page. Please also read through our code of conduct.

Tests

  • Unit tests written thanks to the python module unittest

  • Continuous integration made with Travis (Linux, OSX), and AppVeyor (Windows)

  • Code coverage calculated by the python module codecov

  • The module is ensured to work with Python >= 3.6

  • The module is ensured to work on the platforms Linux, OSX and Windows

  • The script of tests is python/populse_mia/test.py, so the following command launches the tests:

    python3 python/populse_mia/test.py (from populse_mia root folder, for example [populse_install_dir]/populse_mia)

Requirements

  • capsul
  • lark-parser
  • matplotlib
  • mia-processes
  • nibabel
  • nipype
  • pillow
  • populse-db
  • pyqt5
  • python-dateutil
  • pyyaml
  • scikit-image
  • scipy
  • SIP
  • sqlalchemy
  • snakeviz
  • soma_workflow
  • traits

Other packages used

  • copy
  • os
  • six
  • tempfile
  • unittest

GitHub

https://github.com/populse/populse_mia