2010 internship proposal

							                                             2010 internship proposal
    Research team               INRIA Sophia, PULSAR team
    Applicants                  Master students
    Internship duration         5/6 months (February-September 2010)
    Internship location         INRIA Sophia Antipolis, possible visits to IJL Nancy and to CEA Cadarache
    Supervisors                 Dr. Vincent Martin and Dr. Guillaume Charpiat
    Email                       vincent.r.martin@sophia.inria.fr
    Phone                       +33 (0)4 42 25 33 76


Title: dust detection & tracking in fusion plasma by means of fast imaging

Context and objectives:
In the framework of the moniTORE1 project (real-time monitoring of imaging diagnostics applied to Tore
plasma operation), a new activity in dust detection & tracking in fusion by means of fast visible imaging
has started in collaboration with IJL (Institut Jean Lamour, Nancy, France). The ESPRITS team of IJL
works on a small fusion device equipped with a fast imaging camera (~5000 images/sec) able to monitor
the dusts (~some pixels) during plasma discharges. The goal of these experiences is to study the plasma
particle interactions by means of trajectory analysis. To this end, it is necessary to perform a long-term
tracking (~1000 images) of the detected dusts. The INRIA Pulsar team project has a good expertise in
tracking techniques applied to video surveillance of humans or animals subjects. Recently, encouraging
results has been obtained on dust tracking by using particle detection & tracking algorithms [1].
The collaboration between INRIA and IJL aims at developing / improving an automatic detection &
tracking algorithm for dust trajectory analysis in fusion plasma videos provided by IJL. In a second time,
the developed algorithm will be applied to the analysis of data acquired into a Tokamak like ASDEX or
Tore Supra.

Work to be done:
Based on detection & tracking algorithms already developed by INRIA and IJL teams, the work of the
successful candidate will consist in:
        ● adapt the developed algorithms to output useful information on particle and trajectory features
           for further analysis.
        ● proposing a performance evaluation procedure for dust detection & tracking quality estimation,
           algorithm robustness to noise, etc.
        ● adapting a dust detection & tracking algorithm with respect to data and assessment results by
           modifying for instance the spatiotemporal filtering step to improve the signal to noise ratio of
           the videos.

Prerequisite: image & video processing, applied mathematics, Matlab and/or C++ programming

Application: send CV + motivation letter to vincent.r.martin@sophia.inria.fr

References:
[1] I. F. Sbalzarini and P. Koumoutsakos. Feature Point Tracking and Trajectory Analysis for Video Imaging in Cell
Biology, Journal of Structural Biology 151(2):182-195, 2005.

1
    http://www-sop.inria.fr/pulsar/projects/monitore/