Annonce

Les commentaires sont clos.

2010 LADISLAV TAUC Neurobiology Conferences

3 Novembre 2010


Catégorie : Conférence internationale


2010 LADISLAV TAUC Neurobiology Conferences

"From Mathematical Image Analysis to Neurogeometry of the Brain"

16-17 December 2010,       
Salle de La Terrasse - CNRS - Gif sur Yvette - 91190 - France

Organized by UNIC (UPR CNRS 3293: Unit of Neuroscience, Complexity and Information) and GDR MSPC (GDR CNRS 2286: Mathematics of perceptual and cognitive systems).

The conference "From Mathematical Image Analysis to Neurogeometry of the Brain" will be organized in December 2010 in the CNRS Campus at Gif-sur-Yvette (near Paris). This conference is part of the Ladislav Tauc Conferences in Neurobiology series and is jointly organized by UNIC and the GDR CNRS MSPC (Organizers: Yves Frégnac and Gabriel Peyré).

For more information, please visit: http://www.conftauc.cnrs-gif.fr/

---- Abstract ----

Unraveling the complex extraction of geometrical structures in visual scenes is a major bottleneck both in visual neuroscience and artificial vision. The Ladislav Tauc 2010 is focused on this difficult but fascinating issue, and will be a unique forum of interdisciplinary discussions around mathematical image  analysis and neurogeometry. One of the aim of this conference is to illustrate successful attempts in a theoretical field, such as Geometry, to take advantage of the complexity of brain organization to open a common explanatory referential for natural and artificial vision.

On the biological side, the concept of an inner mapping of the outside world, within the layout of the neural circuits responsible for its decoding, is somewhat reminiscent of the assertion of a homeomorphic fit between the world and mind by the American psychologist William James: ‘‘the faculties of our brain are adapted in advance to the features of the world in which we dwell’’. Such a view posits that the topological organization of sensory networks and the morphology of their elements may reflect, at least to a certain degree, constraints in their inner functional organization inherited from the geometrical structure statistics of our visual environment.


The identification of neurally embodied geometrical operators is also crucial to move the research front in mathematical image analysis. Compressing the singular structures of edges and textures in natural images is a key to  obtain high level representations that may be used in a symbolic mode by brains and computers. These efficient representations should be the basis for novel classes of machine learning methods invariant to large image variations.

---- Speakers ----

  • James Bednar (University of Edinburgh, UK)
  • Frédéric Chavane (CNRS, France)
  • Olivier Faugeras (INRIA, France)
  • Jack Gallant (Berkeley, USA)
  • Stephen Grossberg (Boston University, USA)
  • Jan Koenderink (Utrecht University, Netherlands)
  • Stéphane Mallat (Ecole Polytechnique, France)
  • Jean-Michel Morel (ENS Cachan, France)
  • Laurent Perrinet (CNRS, France)
  • Alessandro Sarti (Ecole Polytechnique, France)
  • Eero Simoncelli (NYU, USA)
  • Steven Zucker (Yale, USA)

 

---- Registration ----

Registration is free but mandatory. Register online at: http://www.conftauc.cnrs-gif.fr/