Main Scientific Objectives of the Team
The core expertise of the IMAGES team (the team was previously named TII for Image Processing and Understanding and changed its name in January 2017 to better take into account the evolution of its research topics) is the modeling of images, tri-dimensional and numerical objects, with the development of mathematical models, ranging from the physical acquisition to the high-level interpretation, and artificial intelligence models (spatial reasoning, knowledge representation). The team has also a strong expertise in computer graphics for geometric modeling, image synthesis, virtual reality and 3D interactive systems. The team deals with applications in medical imaging, remote sensing imaging, computational photography and creative industries. The main contributions of the team are described next, and are organized into mathematical methods and artificial intelligence, computer graphics, and applications with a societal impact.
Mathematical methods and artificial intelligence
This section groups our contributions in the field of mathematical modeling for image analysis and understanding. Modeling concerns both data and knowledge about the domain or the data, and is based on various fields of mathematics and computer sciences (statistics, variational approaches, machine learning, algebraic and symbolic artificial intelligence approaches …).
- Acquisition aware image restoration and enhancement
- Stochastic image modeling
- Machine learning and deep learning
- Discrete mathematics, algebraic and structural models, and artificial intelligence
Computer graphics and Virtual Reality
The computer graphics activities of IMAGES mainly take place in the fields of modeling and rendering, with occasional contributions to computer vision. One structuring concept in these activities is the notion of efficient approximations, which helps breaking complexity order when it comes to graphics simulation (e.g., shape deformation, light transport).
- Shape Modeling
- Geometrical and topological analysis
- Rendering
- Vision, virtual reality and 3D printing results
Image processing application to the Society
Almost every theoretical study as presented above find applications with one of our many partners, either institutional (Hospitals, Space centers, …) or industrial. Moreover many of these theoretical studies are gounded in a practical problem presented by one of these partners. It concerns:
- Remote sensing and SAR imagery
- Medical imaging
- Computational photography
- Creative Industries
Fundings (2013-2018)
- ANR: LOGIMA (2012-2016), DESCRIBE (2013-2016), MIRIAM (2014-2018), ALYS (ANR-ASTRID, 2015-2019), ReVeal (2013-2016), STAP (2017-2021), AllegoRI (ANR-LabCom, 2017-2020), iSpaceTime (2010-2014)
- EU: Harvest4D (FET 2013-2016), Reverie (IP 2011-2015), 3DLife (NoE, 2009-2013)
- Joint projects: fundings from Labex DIGICOSME, project Lidex-PIM of IDEX Paris-Saclay, CNES projects, Cifre grants, FSN project funded by BPI (with Ubisoft, CNRS, UPEM and CEA), FUI (G4M)
- Chair: Modélisations des imaginaires (Dassault Systèmes, PSA, Ubisoft, Orange)
- Joint lab with Allegorithmic
Awards (2013-2018)
- IEEE Transactions on Geoscience and Remote sensing best paper award 2016
- 1st PhD prize from Fondation Mines-Télécom (S. Lobry, 2018)
- 2nd PhD prize from Télécom ParisTech (F. Weissgerber, 2017)
- 2nd PhD prize from Télécom ParisTech (E. Guy, 2016)}
- EUSAR best student paper award 2016 (F. Weissgerber)
- Shape Modeling International 2017 Best Paper Award
- Shape Modeling International 2016 Best Paper Award
- Google best student paper, CVMP 2013 (A. Newson)
- Second Place Winner of the Wolfgang Strasser Award 2015 (best paper at
- ACM SIGGRAPH High Performance Graphics)
- IEEE 3DUI 2015 Best Paper Award
- IEEE ICIP 2014 Top 10% Paper Award
- SFIPP 2014 best poster award
- ICPR 2014 Best Scientific Paper Award
- SFGBM best thesis innovation award
- IEEE IPTA 2016 best poster award
- SIAM IS conference best poster award 2018 (A. Houdard)
- CFPT best paper award 2018