The SIGMA team shares a solid mathematical background in statistics and geometry, and is concerned by mathematical objects extracted from physical measurements such as images, sounds, videos, light polarization, printing or chemical sensors. These signals and their properties are processed using either the Bayesian framework, belief function theory for data fusion, information geometry or information theory.

With the common denominator of statistical signal modeling and processing, our research activities belong to scientific communities related to Bayesian inference, evidential decision theory, polarimetric imaging or information forensics and security.

SIGMA maintains strong relationships with important Master Programs such as Data Analysis & Decision making at Centrale Lille, a new master program on machine learning at the University of Lille-3, as well as the statistical signal processing option at Telecom Lille.

Illustrations of our researches :

• Bayesian inference and Monte-Carlo methods : object tracking in video, inverse modeling
• Data fusion : decision using network of sensors
• Information security: intrusion detection and detection of hidden signals, automatic authentication of printed documents
• Polarimetric imaging : detections of artefacts in materials