The “Information Analysis and Management” hub, IAM(https://portal.fli-iam.irisa.fr/home), is tasked with building an infrastructure (hardware and software) accessible to users of national in vivo imaging platforms, in order to facilitate the opening up and reuse of imaging data, and to store, manage and process large sets of clinical and preclinical imaging data and associated metadata.
The hub brings together medical image processing experts from French node and non-node laboratories, under the umbrella of INRIA (Rennes), CNRS (ICUBE, I3S, Creatis, CRMBM), CEA (MIRCen) and INSERM (Brest, Nancy, Grenoble). Scientific coordination of the node has been entrusted to INRIA. Initially coordinated by a tandem formed by Christian BARILLOT and Michel DOJAT assisted by Michael KAIN, the IAM hub has been coordinated by Michel DOJAT with the support of Michael KAIN since Christian BARILLOT passed away in June 2020. We are grateful to our colleague Christian for the work he has accomplished over the years, for his ability to bring people together and for his vision.
The project was implemented in two stages. The first, from 2013 to 2018, was devoted to the creation of a federated image analysis and data management solution that would ensure the interoperability of existing heterogeneous and distributed solutions implementing raw and metadata indexing (e.g., using semantic models or ontologies) of information. Four pre-existing national IAM management solutions were selected: CATI (CEA, Paris-Sud), SHANOIR (INRIA, Rennes), MediBase (CNRS, Strasbourg) and Archimed (INSERM, Nancy). Three existing national solutions for image processing and workflow management were initially selected: VIP (CNRS, Lyon), MedInria (INRIA, Rennes) and BrainVisa (CEA, Paris-Saclay).
The IAM node set up a steering committee to define the node’s action priorities and orientations. A technical manager (Mr. Kain) supervised the developments carried out by a team of up to 15 software engineers based in six different centers. In order to provide the most comprehensive response possible, three working groups were set up:
- The first dealt with interoperability issues, in particular interoperability between data repository solutions.
- The second dealt with image processing problems in order to provide versatile solutions for parallel or massive computing.
- The latter focused on transferring the existing clinical imaging environment and expertise to the emerging preclinical imaging sector.
Since November 2020, the industrialization and operation of the software platform has been entrusted to a temporary consortium of companies, selected following a call for tenders. The contractor offers services based on the data management and processing platform developed by IAM, and the deployment of processing chains on high-performance computing systems to handle large databases.