L2S

• Statistical modeling

We concentrate on time series analysis (long-memory processes, non-stationary processes, robust estimation, missing data); point processes and non gaussian processes; sparse decompositions for source separation; multi-sensor signal processing (source detection, performance bounds, interference cancelling, beamforming optimization).

•  Inverse problems

The group of inverse problems is at the interface between physics, statistics and signal or image processing. The works of the group are based on stochastic models or Markov separable hidden variables in a Bayesian framework, but also sparse approach and convex optimization. Applications are included imaging processing in the broadest sense: microwave tomography imaging, SAR (Synthetic Aperture Radar) imaging, 3D tomography, infrared astronomy imaging, and PET (Positron Emission Tomography) imaging, M/EEG, source separation etc.