EN FR
EN FR




Section: New Results

National Initiatives

  • SOFA-InterMedS (2009–) Participants: R. Anxionnat, M.O. Berger, E. Kerrien, A. Yureidini.

    The SOFA-InterMedS large-scale INRIA initiative is a research-oriented collaboration across several INRIA project-teams, international research groups and clinical partners. Its main objective is to leverage specific competences available in each team to further develop the multidisciplinary field of Medical Simulation research. Our action within the initiative takes place in close collaboration with both Shacra INRIA project-team in Lille and the Department of diagnostic and therapeutic interventional neuroradiology of Nancy University Hospital. We aim at providing in-vivo models of the patient's organs, and in particular a precise geometric model of the arterial wall. Such a model is used by Shacra team to simulate the coil deployment within an intracranial aneurysm. The associated medical team in Nancy, and in particular our external collaborator René Anxionnat, is in charge of validating our results.

  • ANR ARTIS (2009-2012)

    Participants: M.O. Berger, A. Eryildirim, E. Kerrien.

    The main objective of this fundamental research project is to develop inversion tools and to design and implement methods that allow for the production of augmented speech from the speech sound signal alone or with video images of the speaker's face. The Magrit team is especially concerned with the development of procedures allowing for the automatic construction of a speaker's model from various imaging modalities.

  • ANR Visac (2009-2012)

    Participants: M.O. Berger, B. Wrobel-Dautcourt.

    The ANR Visac is about acoustic-visual speech synthesis by bimodal concatenation. The major challenge of this project is to perform speech synthesis with its acoustic and visible components simultaneously. Within this project, the role of the Magrit team is twofold. One of them is to build a stereovision system able to record synchronized audio-visual sequences at a high frame rate. Second, a highly realistic dense animation of the head must be produced.