WP5: Data, brain and vision

A substantial amount of research will be devoted to the development of new tools for data analysis inspired by brain processes. We will extend existing models of cortical organization to geometrical structures accounting for other higher dimensional fibers. In particular subriemannian, contact and symplectic structures will be considered. Stochastic propagation of the visual signal will be studied to describe both eye movements and propagation along long range connectivities in low level areas. We will also develop a new approach for the problem of learning group invariances and extra features from data. The whole work will lead to a theory of vision and image analysis in Lie groups, inspired by the functionality of the visual cortex. Instruments of numerical analysis wll be introduced to implement the proposed models.