The recruited PhD student will focus on studying and implementing machine learning techniques applied to pattern, colors and luminance recognition. The developed techniques will be applied to the field of automatic recognition of the quality of Tahitian pearls. The machine learning techniques shall use mathematical descriptors computed from (i) photos of the surface of the beads, (ii) X-ray images and (iii) acoustic signatures. Note that some of these mathematical descriptors have already been defined in particular as regards the color of beads (see publications below). The automatic classification results obtained by the machine learning algorithm will be compared with the classification performed by human experts. This comparison will serve to adjust and validate the implemented machine learning procedures. This work will be conducted as part of the RAPA project (seehttps://sites.google.com/site/rapaproject/home) supported by the French Ministry of Overseas. The project will be conducted in the building of the GePaSud Laboratory at the University of French Polynesia in Tahiti.

Student Profile : The applicant should hold a MSc in Computer Science and be specialized in 1) machine learning with knowledge in 2) image processing and/or 3) acoustical imaging.Being able to write in English is essential.

keywords : Machine learning, fusion, classification, image processing, acoustics.

Funding : Doctoral grant (obtained) from the MESR (French Ministry of Research, see http://fr.wikipedia.org/wiki/Contrat_doctoral)

Application Deadline : see https://sites.google.com/site/rapaproject/home/job-offers

Research Director : Pr Alban Gabillon
Tel : (+689) 40 803 880
email : alban.gabillon@upf.pf
Co-Research Director: Dr. Jean-Martial Mari
Tel : (+689) 40 866 439
email : jean-martial.mari@upf.pf
Supervisor : Dr. Sébastien chabrier
Tel : (+689) 40 803 805
email : sebastien.chabrier@upf.pf
Place : Laboratoire GePaSud, Bâtiment OGT, Université de la Polynésie Française, Tahiti.

Publications:
Martin Loesdau, Sébastien Chabrier, Alban Gabillon. Automatic Classification of Tahitian Pearls. Proceedings of the 6th Image Processing & Communication (IPC) conference. September 2014. Bydgoszcz. Poland. Advances in Intelligent Systems and Computing 313. Image Processing & Communications Challenges 6. Springer. ISSN 2194-5357.

Martin Loesdau, Sébastien Chabrier, Alban Gabillon. Hue and Saturation in the RGB Color Space. Proceedings of the 6th International Conference, ICISP 2014, Cherbourg, France, June 30 – July 2, 2014. Lecture Notes in Computer Science 8509. pp 203-212. ISBN 978-3-319-07997-4.