Vous êtes ici

VISAGES TEAM : Post-doc MUSIC “Detection and segmentation framework for multiple sclerosis lesions: Application to a large clinical multi-centric database”

Lieu: 
Contexte: 

Environment:

This post-doctoral position will take place at Inria/IRISA, UMR CNRS 6074, in the VisAGeS U746 research team. The work will be conducted in close link with the MRI experimental platform at Neurinfo (http://www.neurinfo.org) and the neurologists and radiologists involved in the project.

Context:

VisAGeS U746 is a research team from Rennes 1 University, jointly affiliated with Inserm and Inria. It is also part of IRISA (UMR CNRS 6074) and is located in Rennes, France on both medical and science campuses. The objective of the team is to work jointly with clinicians, radiologists from the university hospital to propose new advances in medical image processing. Among other applications, the team is involved in multiple sclerosis (MS) image processing studies, in close collaboration with neurologists and radiologists. MS is a frequent neurological, inflammatory and demyelinating, disease affecting young adults, and is a source of several disabilities, including ambulatory.

Magnetic resonance imaging has emerged as a powerful noninvasive tool of diagnosis, description of the natural history of the disease and treatment monitoring of MS. In addition, MRI findings have been used to explore drug efficacy in clinical trials. Conventional MRI surrogates provide information at the macroscopic level but lack sensitivity and specificity in identifying the full extent of underlying MS pathology. They also show relatively weak relationships to clinical status such as predictive strength for clinical change (called as the clinical-MRI paradox). With the advent of disease modifying drugs, there is a need for robust and specific MR markers to characterize the pathology. One of the most current relevant markers in the clinical follow-up is to characterize multiple sclerosis lesions (MSL) and their evolution overtime.

Mission: 

Medical image computing in MS requires the application of image processing workflows. The goal of this post-doctoral position is to provide and combine a range of algorithms to detect, segment and follow overtime the MSL robustly enough to work on a large clinical database. It will provide to the community a combination of efficient methods for medical image analysis in MS. The combination of methods will be calibrated and evaluated based on a sizeable amount of training and test images with high quality segmentations from multiple experts. Once the algorithms will be combined, they will be integrated in a production workflow that will be used by the clinical health network MUSIC that is covering the western part of France. This work will rely on methods developed in the research team, among others [1-3].

The post-doctorate will further use all these tools to process the MUSIC database in order to evaluate the focal lesion burden as well as their evolution in time.

From a methodological perspective, this work will deal with registration between modalities, time points and patients segmentation of brain tissues and MS lesions preprocessing of images (denoising, bias correction, intensity normalization, …) statistical comparison and fusion of individual MSL segmentation algorithms using label fusion methods.

Profil / compétences: 

This work will require strong knowledge in the fields of applied mathematics (statistics, optimization), and image processing (image segmentation, registration…). A PhD thesis in one of those fields will thus be required. A good knowledge of computer science tools will also be required, especially in object oriented programming (C++), Matlab or python.

Diplôme requis: 
PhD thesis
Lieu de travail: 
Rennes
Durée du contrat (en mois): 
12
Quotité: 
100%
Corps / catégorie: 
Post-doctorant
Salaire Brut / Mens €: 
around 2100 € net / month
Date prévisionnelle d'embauche: 
as soon as possible, for 1 year with possible renewal
Candidater: 

Olivier Commowick (olivier.commowick@inria.fr)
Christian Barillot (Christian.Barillot@irisa.fr)
Anne Kerbrat (anne.kerbrat@chu-rennes.fr)

Link to the full position description:  https://team.inria.fr/visages/files/2011/07/PostDoc_MUSIC_en.pdf