G e o S t a t
Geometry & Statistics in acquisition data
PDF Print E-mail

GeoStat is an INRIA project located at INRIA Bordeaux Sud-Ouest (INRIA BSO), inside the theme: applied mathematical computation and simulation, optimization, learning and statistical methods. The project makes fundamental and applied reasearch on new nonlinear methods for the analysis of complex systems, turbulent and natural signals, using paradigms and tools coming from the notions of scale invariance, predictability, universality classes and nonlinear physics. We study the parameters related to common statistical organization in different complex signals and systems, we derive new types of sparse and compact representations, and we develop tools for the analysis of complex signals that better match the statistical and geometrical organisation inside these data: as a typical example, we cite the evaluation of cascading properties of physical variables inside complex signals through optimal wavelets, resulting in radically new methods for evaluating complex signals' dynamics.

With our partners at LEGOS (CNRS UMR 55 66, Toulouse) and ICM-CSIC (Barcelona, Spain), we develop a set of tools for studying common statistical organizations in different complex signals, based on the precise computation of singularity exponents, contemplated in a microcanonical formulation; this microcanonical formulation allows for the computation of quantities without ergodic hypothesis, it give access to the geometrical organization of the multiscale hierarchy present in complex signals without the consideration of grand ensembles and stationarity, and it paves the way for effective nonlinear approaches for inference and classification in complex signals. This approach leads also to original sparse and compact representations, and is associated to specific reconstruction formulae.

Since February 2014, GEOSTAT is an associated team with India IIT Roorkee's team of Prof. D. Singh. Link to associated team "OPTIC" web page .

GeoStat's research thematics are centered on the following theoretical developments:

  • Nonlinear signal processing using methods from complex systems, statistical physics and nonlinear physics,
  • Sparse and compact representations, signal reconstruction,
  • Predictability in complex systems,
  • Optimal wavelet decomposition,
  • Analysis, classification, detection in complex signals.

and the following applied objectives:

  • Analysis of complex and turbulent signals in earth observation and remote sensing,
  • Nonlinear approaches to the analysis of heartbeat signals.
  • Analysis of atmospheric turbulence in optical imaging (nonlinear phase reconstruction in Adaptive Optics).
  • Speech analysis.

GeoStat is working in close collaboration with the following teams:

GeoStat is a member of GDR PHENIX.

GeoStat is a member of GDR ISIS.

GeoStat is a member of GDR AMF.

Last Updated on Wednesday, 22 February 2017 15:37