Image analysis platform

The image analysis platform enables researchers to work together by sharing medical imaging data as well as image-analysis pipelines. The medical images can be stored in a central XNAT database.

What is BBMRI-NL’s Image analysis platform? 

The Image Analysis Platform facilitates running medical image-analysis pipelines on medical images. It provides the user with an XNAT medical image archive for storing image data, a flexible platform for  running analysis pipelines on these images to produce imaging biomarkers, and a viewer to inspect intermediate and final results. Image analysis pipelines can be shared for running at your own institute or on national supercomputing facilities. The tools used for processing are provided under an open-source license.

The image below gives an overview of the components :

  • Study Governor: The Study Governor is an automated data and state manager. This service keeps track of the status of all data samples (usually scan sessions) in the study and automates the process. This is the brain of the infrastructure. It exposes a REST API and a web interface to monitor and manage the study process. (Apache 2.0)
  • Taskmanager: Taskmanager manages tasks for manual interaction with image and derived data. Tasks can be added to the Taskmanager by the Study Governor via its REST API but can also be used stand-alone.
  • XNAT: XNAT is an open source imaging informatics platform developed by the Neuroinformatics Research Group at Washington University. It facilitates common management, productivity, and quality assurance tasks for imaging and associated data. XNAT is a hosted national service by Health-RI TraIT, but it can also be hosted in your institute.
  • ViewR: Customizable viewing software based on MeVisLab. This viewer receives its layout from the Taskmanager using a templating system and lets the user perform ratings, annotations or inspections on medical images in a coordinated fashion. A more lightweight alternative for the ViewR is pysnap. Please inquire with the platform coordinator if you are interested in using the VIewR.
  • FASTR: FASTR is a framework that helps creating workflows of different tools. The workflows created in FASTR are automatically enhanced with flexible data input/output, execution options (local, cluster, etc) and solid provenance. This tool is developed by the Erasmus MC in BBMRI WP3 and WP5. (Apache 2.0)
  • PIM: Web framework for online Pipeline Inspection and Monitoring. Exposes a REST API for being interoperable with pipeline engines. Fastr supports visualizing the pipelines and their progress in PIM. This tool is developed by LUMC in BBMRI WP3. (GPL)
  • StrongR: enables truly elastic scaling of batch processes in the cloud, allowing pipeline frameworks to exploit the benefits of cloud computing. It will create worker nodes when processing power is required, and shut them down when not. It allows process upscaling when there is more demand for computational power. StrongR can interface with multiple cloud providers, allowing users to do processing in the cloud of their choice. StrongR is developed by Erasmus MC in BBMRI WP3. (Apache 2.0)

Why BBMRI-NL’s image analysis platform?

Biomedical imaging plays a key role in health research. The imaging platform facilitates the analysis of medical imaging data and generation of imaging biomarkers for research. The platform is easy to deploy either in your local institute or on national research infrastructure. The provided image-analysis pipelines are established methods that have proofed their performance on large datasets.

Intended users

The platform is developed to enable researchers to extract biomarkers from imaging data. Imaging analysis researchers that offer their image analysis pipelines also benefit from using this infrastructure.

Implemented pipelines

The Image Analysis platform is constantly being developed.

The following image analysis pipelines have been implemented in the platform:

  • Population based statistical analysis of brain structures (LUMC)
  • Whole brain segmentation (EMC)
  • Tissue segmentation (EMC)
  • White-matter-lesion segmentation (EMC)
  • Brain-structure segmentation (EMC)
  • Epi-cardial-fat segmentation (EMC)
  • Intracranial Volume (UMCU)
  • DTI measures (UMCU)
  • Subcortical segmentation (Donders Institute)
  • Independent Connectivity Parcellation (Donders Institute)
  • Spatial Trend Surface Models (Donders Institute)

For information to start using BBMRI-Imaging

Contact the Imaging Platform Coordinator: Marcel Koek or the TraIT Service Desk

References and links

Website of the European Population Imaging Infrastructure

Description of  the XNAT service as hosted by Lygature TraIT

TraIT Service Desk

Organization on GitLab were platform specific source code is located

BBMRI-NL’s WP leaders:

Aad van der Lugt:

Wiro Niessen:

Hilleke Hulshoff Pol: