I hold a full time position at GE Healthcare since 2003. I am currently senior research engineer in this company, in charge of developping image processing applications for interventional cardiology. I am simultaneously preparing a PhD at Universite Paris Est (UPE), France, since December 2009.
My PhD activities take place at ESIEE in the Computer Science Department, and in the A3SI team of the Laboratoire d'Informatique Gaspard Monge (LIGM) under the supervision of my PhD advisors Laurent Najman and Régis Vaillant.
My PhD thesis deals with image processing applications for interventional cardiology, in the specific context of stenting (video). These procedures are monitored under Xray fluoroscopy delivering real-time imaging to the clinician to guide his gesture (video). In the Xray images the clinician can observe the location and motion of his tools with respect to the patient's anatomy. However, the finest ones, namely guide-wires and stents, are difficult to visualize due to their low signal to noise ratios. The objective of my work is to provide means to enhance their visibility.
In the presented work a wide variety of image processing techniques (segmentation, detection, tracking, registration...) are put together to build efficient and thoroughly validated image processing algorithms.
Stent visualization enhancement relies on detecting, tracking and segmenting the tool supporting the stent.
In a second step, image registration, temporal integration and dedicated processing allow to produce an enhanced stent image (see
video example on a clinical case).
(For more examples see StentViz for instance). A detailled analysis of the problem and a fully automated algorithm are presented. Several extensions that aim at improving its performances and scope of applicability have been investigated.
Guide-wire enhancement falls into the field of curvilinear structure segmentation. A comparison of state of the art techniques is presented and a new technique is proposed.