Post doctoral position 
Topic : Adaptative prior in level-sets segmentation

The position is offered at LIAMA (Beijing), in the team RSIU. 
The candidate will take part of a dynamic, talentuous and international team, combining an exciting research work and a rich cultural environment. 

The starting date will be at the end of summer 2007, depending on the availability of the selected candidate, for 12 months.
The salary will be around 2,000¢ã gross per month.

Conditions of application and full subject description, see: http://www.inria.fr/travailler/opportunites/postdoc/postdoc.en.html
The deadline for application is July 2007.

Description
Shape-prior active contours are powerful models for the segmentation of objects in the image. Level-sets approaches in particular have been widely used in medical or remote sensing applications. However, the drawback of the introduction of a shape prior constraint is that  it compells the contour to fit with this prior (or an affine transformation of it): it cannot handle cases of an  erroneous shape knowledge.  The aim of this research work is to propose a model that will take into account an adaptative prior shape for object segmentation from level-sets  active contour models. This would significantly improve  the  accuracy of the segmentation.  A parallel processing  will enable the simultaneous segmentation of several independant objects and reduce  computation time.  The applied objective is to automate the task of digital map updating using optical high resolution images ("GoogleMap-like" images). The objects of interest are the buildings, the prior shape being given by exogeneous GIS data. 

The candidate will have a PhD in CV, PR or related domains. She/he will have solid knowledge in theory and application of level-sets. A strong background in applied mathematics and numerical analysis  is required. The programmation will be done in C++.
The candidate will be fluid in english.

The position is offered at LIAMA (Beijing), in the team RSIU. RSIU's  aim is to develop and validate new mathematical models and computational algorithms for the analysis of deformation, change, or motion observed from multi-temporal data set, particularly from satellite images and geographical data.  

Application can be done on line from http://www.inria.fr/travailler/opportunites/postdoc/postdoc.en.html
 Or 
by sending a full CV, together with your 2-3
most relevant publications and 2 recommendation letters to  prinet@nlpr.ia.ac.cn .