Skip to Main content Skip to Navigation
Habilitation à diriger des recherches

Image processing methods for 3D intraoperative ultrasound

Abstract : This document constitutes a synthesis in preparation for my habilitation degree in computer science. I am now researcher at INRIA Rennes since September 2001. From September 2001 to January 2004, I was a researcher in the Vista project headed by Patrick Bouthemy, and moved to the Visages project headed by Christian Barillot. In January 2010, I moved to the Serpico team headed by Charles Kervrann that focuses on ”imaging and modeling intracellular dy- namics of molecular architectures”. This document presents part of my work among the Visages team. Actually, this habilitation thesis will focus on image processing aspects of intraoperative ultrasound in neurosurgery. My work on non-rigid registration will not be described here. The work on non-rigid registration began during my PhD thesis, where the three main contributions were the design of a 3D non-rigid registration method based on optical flow [72], the incorporation of local constraints [74] and the retrospec- tive evaluation of inter-subject registration [71]. I continued working on image registration, with Anne Cuzol and Etienne M ́emin using fluid motion descrip- tion [44], with Nicolas Courty on GPU accelerated registration [42] and on evaluation of non-rigid registration techniques: with Mallar Chakravarty and co-authors [29] for deep-brain stimulation planning; with Arno Klein and co- authors [92] concerning inter-subject brain registration. In the last decade, it has become increasingly common to use image-guided navigating systems to assist surgical procedures [51]. The reported benefits are improved accuracy, reduced intervention time, improved quality of life, reduced morbidity (and perhaps mortality), reduced intensive care and reduced hospital costs. Image-guided systems can help the surgeon plan the operation and provide accurate information about the anatomy during the intervention. Image-guided systems are also useful for minimally invasive surgery, since the intraoperative images can be used interactively as a guide. Current surgical procedures rely on complex preoperative planning, includ- ing various multimodal examinations: anatomical, vascular, functional explo- rations for brain surgery. Once all information has been merged, it can be used for navigation in the operating theatre (OR) using image-guided surgery systems. Image-guided surgery involves the rigid registration of the patient's body with the preoperative data. With an optical tracking system, and Light Emitting Diodes (LED), it is possible to track the patient's body, the micro- scope and the surgical instruments in real time. The preoperative data can then be merged with the surgical field of view displayed in the microscope. This fusion is called “augmented reality”. Unfortunately, the assumption of a rigid registration between the patient's body and the preoperative images only holds at the beginning of the procedure. This is because soft tissues tend to deform during the intervention. This is a common problem in many image-guided interventions, the particular case of neurosurgical procedures can be considered as a representative case. When dealing with neurosurgery, his phenomenon is called “brain shift”. Although the impact and the magnitude of soft tissue motion have been studied over the last few years, this phenomenon is still poorly understood. Soft tissue deformation can be explained by physiological (steroids, diuretic medication, mechanical ventilation) and mechanical factors (CSF leakage, pa- tient positioning, tumor nature and location, craniotomy size, gravity [117], etc). The magnitude of brain shift shows striking differences at each stage of surgery. Brain shift must be considered as a spatio-temporal phenomenon, and should be estimated continuously, or at least at key moments, to update the preoperative planning. To do so, one possibility is to deform the anatomical and functional images according to the estimated deformation.
Complete list of metadatas

Cited literature [240 references]  Display  Hide  Download
Contributor : Pierre Hellier <>
Submitted on : Monday, November 15, 2010 - 2:28:01 PM
Last modification on : Thursday, March 5, 2020 - 5:46:02 PM
Document(s) archivé(s) le : Friday, December 2, 2016 - 10:26:26 PM



  • HAL Id : tel-00536133, version 1



Pierre Hellier. Image processing methods for 3D intraoperative ultrasound. Modeling and Simulation. Université Rennes 1, 2010. ⟨tel-00536133⟩



Record views


Files downloads