Summary
This project will develop and validate an accurate and robust 3D motion vector estimation technique using medical ultrasound. A 3D motion vector estimation approach based on extension to 3D of a technique developed in our research unit based on transverse oscillations (TO) is already under development. It combines original image formation and post processing methods. The project will contribute to: the development of the 3D TO image formation and 3D motion estimation techniques adapted to TO, the demonstration and quantification of its potential for cardiac imaging. Simulation of healthy and pathological cases will be produced to assess the potential of the technique for early detection of cardiovascular pathologies. Once the method has been validated using simulations, the project will undertake experimental validation of the approach using a research ultrasound scanner and data acquired from a physical phantom. A final feasibility study will be then performed on healthy volunteers.
Description
Cardiovascular pathologies remain one of the leading causes of mortality in western countries. Among all means to diagnose, quantify and follow these pathologies, medical ultrasound (US) imaging is probably the one having most assets: it is fast, safe, transportable, and cheap compared to other imaging modalities. As a result, any progress in that modality will have a major impact on the population. Unfortunately US imaging has also some drawbacks: it is user dependent and most of the commercial systems only provide high quality 2 dimensional images, whereas the organs under investigation are 3D. This project will contribute to the development of US based imaging modules for 3D echocardiography. In this context we are particularly interested in the development of high quality (accurate and robust) 3D motion vector estimation methods. We have previously shown that a specific image formation technique called transverse oscillations (TO) is particularly well adapted to motion detection and that the combination of such images with phase based algorithms is particularly accurate and robust. The transfer to ultrafast 3D US in already under study and the recruited PhD candidate will be part of an already quite mature project. An extremely important task will be to develop the simulations and experiments to evaluate and quantify the accuracy of the methods developed in realistic simulations and experiments corresponding to both healthy and pathologic situations. These results should also highlight limitations of the technique, which the PhD candidate will then model and for which he/she will develop improvements. The project will be done in an exciting and stimulating environment since it is part of an European Training Network (ETN) financed by the European Commission. The ETN aims at training the European researchers of the future. The recruited PhD candidate will be exposed to the different sectors of biomedical research namely academia, industry and clinics. Three summer schools and closing conference will be organized in the different partnerβs institutions with the aim of reinforcing the network and developing both the intellectual and human qualities of the PhD candidate. The profile of the recruited PhD candidate should include a strong background in signal/image processing and programming (electrical engineering / computer science) and a strong predilection to work on applied research close to both industrial and clinical transfer. Mobility is a mandatory since secondment periods abroad are expected during the project. As a consequence adaptability is also a key quality.
Specific requirement: The candidates must not have resided in France for more than 12 months in the 3 years immediately prior to the recruitment date, and not have carried out their main activity (work, studies, etc.) in that country.
Application by e-mail: denis.friboulet@creatis.insa-lyon.fr and herve.liebgott@creatis.insa-lyon.fr