|
Postdoc, INRIA Sophia Antipolis / CNES
Keywords: Marked point processes, Parameter Estimation, Object detection Contact: |
 |
Abstract:
My work at INRIA concerns object detection and tracking on a sequence of high resolution optical images acquired either by airplanes or by high resolution satellites. The first part aims at developing a parameter estimation method for marked point processes which is both accurate and fast, and possibly parallelizable. The second part concerns the application to object change detection for harbor activity monitoring.
Short Bio:
Aurélie Boisbunon received a master degree in mathematical engineering from the National Institute of Applied Sciences (INSA) of Rouen, France, in 2008 and a Ph.D. degree in mathematics from the University of Rouen, France, in 2013.
In 2013 she has been working for 6 months as a Postdoctoral Fellow at University of Tokyo, Japan. Currently she is working as a Postdoctoral Fellow with AYIN team, INRIA Sophia Antipolis, France. Her main research interest lies in the field of statistical image analysis and processing. In particular, her work concerns statistical modeling, object detection and tracking on a sequence of high resolution optical images.
Last publications:
A. Boisbunon, Y. Maruyama. “A predictive Stein’s effect even in the low-dimensional case”, submitted 2013.
A. Boisbunon, S. Canu, D. Fourdrinier, W.E. Strawderman, M.T. Wells. “AIC and Cp as estimators of loss for spherically symmetric distributions”, submitted 2013.
A. Boisbunon, S. Canu, D. Fourdrinier. “Criteria for variable selection with dependence”, Workshop on New Frontiers in Model Order Selection, NIPS 2011.
A. Boisbunon, S. Canu, D. Fourdrinier. “Critères robustes de sélection de variables pour le modèle linéaire via l’estimation de coût”, 18èmes Rencontres de la Société Francophone de Classification. September, 2011.
