PhD FELLOWSHIP AT INSTITUT TELECOM SUDPARIS, FRANCE

POSITION : PhD fellowship in Medical Image Analysis
ORGANIZATION : Institut TELECOM SudParis, France
LOCATION : INSTITUT TELECOM SUDPARIS, FRANCE

Description :

Three-Year Ph.D. Fellowship in Medical Image Analysis at Institut TELECOM / TELECOM & Management SudParis, France (www.institut-telecom.fr / www.it-sudparis.eu)

LABORATORY:

ARTEMIS Department (www-artemis.int-evry.fr)

RESEARCH SUBJECT:

Spatio-temporal statistical variational models for the quantitative assessment of myocardial perfusion in MRI.

DESCRIPTION:

This doctoral research aims at elaborating non supervised computer-aided diagnosis (CAD) tools for the quantitative assessment of myocardial perfusion in Magnetic Resonance Imaging (MRI). First-pass perfusion MRI is widely used for diagnosing cardiovascular pathologies and evaluating their physiological impact on heart tissues. Myocardial perfusion is assessed by monitoring the MR signal enhancement during the first-pass transit of a vascular contrast agent through the cardiac chambers and muscle. To this end, the heart is partitioned into a set of anatomical segments according to a standardized subdivision model of the myocardium. For each segment, a so-called first-pass curve, defined as the regional average MR signal value as a function of time, is then computed from which a set of clinically-relevant parameters are finally derived. Quantitative perfusion assessment is achieved by comparing the estimated parameters to normative values. Automated perfusion assessment is faced with the complex problem of segmenting the myocardium in image sequences with high nonlinear contrast variations (with possible pathology-induced alterations) and non rigid motion artefacts originating from respiration and cardiac deformations. Solving this issue will be the major scientific objective of this Ph.D. Previous research at ARTEMIS has led to elaborating a purely spatial methodology relying on statistical variational techniques for non rigid image registration and deformable model-based segmentation. This thesis will further investigate and extend these ideas to efficiently exploit the spatiotemporal relationships between MR data at the pixel, organ and measurement scales. In a second stage, research will be directed towards improving segmentation performance by using statistical prior knowledge on the shape, deformations and contrast enhancement of cardiac structures. Extensive validation of the resulting methodology will be performed in the framework of a close partnership with the Radiology Department of Pitié-Salpétrière Hospital in Paris.

REQUIREMENTS:

The successful candidate should have a background in applied mathematics, electrical engineering or computer science. He/she should hold a MSc. Degree (or a French Master Degree). Engineering degrees will be also considered provided that the applicant can demonstrate a first research experience. Good programming (preferably C/C++) and mathematical skills are essential. Experience in any such fields as medical image analysis, image processing, information theory, statistics or numerical analysis, will be considered favourably. Written and spoken English is necessary. The candidate will be available to begin his/her studies in November 2008 or later. Criteria on nationality and age do not apply. The application of women candidates is encouraged.

SALARY:

Fully-funded 3 year Ph.D. fellowship of Institut TELECOM. The net income will be about 1400 per month. (16,800 per annum).

APPLICATION:

Send a full CV, a cover letter and the names and contact information of two referees to:
Dr Nicolas Rougon (Nicolas.Rougon@it-sudparis.eu)

Contact address :
Dr. Nicolas Rougon
Associate Professor
ARTEMIS Department
Nicolas.Rougon@it-sudparis.eu

Keywords :
Medical imaging, image analysis, variational and PDE techniques, computer-aided diagnosis, myocardial perfusion, magnetic resonance imaging.

Name : Prof. Nicolas ROUGON
Email : Nicolas.Rougon@it-sudparis.eu