ARIANA (INRIA)

INRIA

Ariana is a joint research group created by INRIA, CNRS and UNSA (University of Nice-Sophia Antipolis) in 1998 and located at INRIA Sophia-Antipolis. It currently comprises 4 staff members (among them one is part-time because he is also the director of INRIA Sophia-Antipolis research centre), 3 research collaborators, 2 invited professors, 5 PhD students, 3 Master's students (DEA) and 3 interns from Ecole Polytechnique and from Ecole Normale Supérieure.

The Ariana research group aims to pool theoretical mathematical tools, either probabilistic or variational, and test them on inverse problems derived from concrete applications in Earth observation and cartography. This in no way involves juxtaposing the studies in both areas, but rather means studying the ties between these two families of methods, in particular, in their hierarchical / multigrid aspects and in estimation of parameters. The principal research topics of the Ariana reearch group include: Markovian modelling, variational approaches, and probabilistic geometry. The study of MCMC methods is of particular interest for Ariana. Such sampling techniques have already been used for parameter estimation in a restoration problem on real aerial and satellite images (in cooperation with the French Space Agency (CNES) and Alcatel Space Industries) and for road detection on aerial images using probabilistic geometry (collaboration with the French National Mapping Institute (IGN)).

The role of Ariana within MOUMIR is to combine its expertise with that of the University of Cambridge and of TCD in order to set up a new multi-resolution method for segmentation of textured and non textured images into objects, based on Bayesian theory and Complex Wavelet Decomposition. In addition, research work is to be conducted in parallel with BGU on the characterization of textures by means of parametric models using the 2D Wold decomposition in order to achieve automatic parameterization for indexing.

Josiane Zerubia has been a permanent research scientist at INRIA since 1989 and director of research since July 1995. She was head of a remote sensing laboratory (PASTIS, INRIA Sophia Antipolis) from mid-1995 to 1997. Since January 1998, she has been in charge of a new research group working on remote sensing (Ariana). Before this, she was with the Signal and Image Processing Institute at USC, Los Angeles as a post-doctoral researcher. She also worked as a researcher for LASSY (University of Nice and CNRS) from 1984 to 1988 and in the Research Lab. of Hewlett Packard in France and in Palo Alto (CA) from 1982 to 1984. She took a Doctor of Engineering degree in 1986; a PhD in 1988 and an ``Habilitation'' in 1994. She has been part of the IEEE IMDSP Technical Committee (SP Society) since 1997 and associate editor of IEEE Transactions on Image Processing since 1998. She has also been a member of the editorial board of the French Society for Photogrammetry and Remote Sensing (SFPT) since 1998. Her current interests are image processing using probabilistic models and neural networks, parameter estimation and optimization techniques.

Ian Jermyn received the PhD degree in theoretical physics from the University of Manchester, U.K., in 1991. After working for a total of three years at the International Centre for Theoretical Physics in Trieste, Italy, he began study for a PhD in computer vision in the Computer Science department of the Courant Institute of Mathematical Sciences in New York, receiving the PhD in July 2000. He joined the MOUMIR project as a postdoc in the Ariana research group at INRIA Sophia-Antipolis in August 2000. Since September 2001, he has been a permanent researcher at INRIA.

Roberto Cossu received the "Laurea" (M.S.) degree in electronic engineering (summa cum laude) from the University of Genoa, Italy, in 1999, and the PhD in Image Processing and Recognition from the University of Trento, Italy, in 2003. During his PhD he was a member of the Remote Sensing Laboratory in the Department of Information and Communication Technology, University of Trento. In April 2003, he joined the EU project MOUMIR and since then has been working as a postdoc in the Ariana research group at INRIA Sophia Antipolis, France. His main research contributions are in the area of remote-sensing image processing and recognition. In particular, his interests include: neural networks for classification purposes, partially unsupervised updating of land-cover maps, supervised and unsupervised change detection techniques, texture segmentation. In these subject areas, he has carried out research within the framework of several national and international projects.

Karen Brady received a BA in Mathematics from Trinity College Dublin in 1998 and a MSc in Operations Research from The George Washington University in 2000. She joined the Ariana team at INRIA in October 2000 as a MOUMIR researcher and will study toward a PhD degree in image processing.