Medical Image Processing

Medical images are a fundamental element in medical diagnosis and treatment, as they reveal the internal anatomy of patients. At Vanderbilt, research in this field primarily focuses on image segmentation, image registration and image-guided surgery.

Medical Imaging Processing

Benoit Dawant

Cornelius Vanderbilt Professor of Engineering
Professor of Electrical Engineering
Director of the Vanderbilt Initiative in Surgery and Engineering

Benoit Dawant works at the interface between engineering and medicine with particular expertise in image processing and surgical guidance. Current projects focus on the development and clinical evaluation of algorithms and systems to assist in the placement of deep brain stimulators used for the treatment of Parkinson's disease and other movement disorders, the placement and programming of cochlear implants used to treat hearing disorders, the removal of brain tumors and the creation of radiation therapy plans.

For more than 10 years and with National Institute of Health funding, Dawant and Vanderbilt researchers have developed a system that facilitates the pre-operative, intraoperative and post-operative phases of deep brain Medical Image Processing stimulation (DBS) procedures. The system, which consists of a data repository called CranialVault and a suite of software tools called Cranial Vault Explorer, is the first computer-assistance system that spans the entire spectrum of the procedure.

In addition to his research, Dawant is at the helm of the Vanderbilt Initiative in Surgery and Engineering, an interdisciplinary, trans-institutional center whose mission is the creation, development, implementation, clinical evaluation and commercialization of methods, devices, algorithms and systems designed to facilitate interventional processes and their outcome. The center facilitates the exchange of ideas between physicians, engineers and computer cientists. Primarily funded by federal research support, the center is also engaged with industrial partners for the commercialization of its intellectual property, the early evaluation of industrial devices and techniques and the joint development of innovative solutions.

Bennett Landman

Assistant Professor of Electrical Engineering

Bennett Landman's projects range from understanding the neurological basis of psychological disorders and mapping brain tumors to statistical method development and visualizing abdominal defects. The common theme that unifies his work is capturing quantitative information from three (or higher) dimensional medical images.

Under the Vanderbilt University Institute of Imaging Science, he leads the Center for Computational Imaging, which develops support and informatics tools and infrastructure to help advance imaging research for the center's users, collaborators and the medical imaging community. The group aims to provide infrastructure, image analysis and informatics expertise to facilitate basic science research using imaging data, while developing novel approaches to computational imaging techniques. Landman is currently working on projects that investigate statistical label fusion techniques and multimodal MRI approaches.

In addition to this work, Landman is developing computational tools for imaging genetics analysis under National
Institute of Health funding. The emerging field of imaging genetics combines modern statistical genetics methods, which embrace quantitative phenotyping, with neuroimaging, which is capable of providing detailed structural and functional phenotypical measurements.

Jack Noble

Research Assistant Professor of Electrical Engineering

Longtime cochlear implant (CI) users are reporting dramatic improvements in their hearing, thanks to new imageguided programming methods developed by Vanderbilt University researchers. Jack Noble is collaborating with an interdisciplinary team that includes Benoit Dawant, René Gifford and Robert Labadie to develop software that can boost the performance of existing cochlear implant technology and improve the quality of hearing for CI recipients.

Current devices use a combination of surgically implanted electrodes that stimulate auditory nerve pathways and an external sound processor worn behind the ear to provide hearing sensations. The new automatic technique uses patients' pre- and post-implantation CT scans to determine the location of the implanted electrod Medical Image Processing es and where stimulation overlap is occurring, possibly causing interference in the transmission of signals. The image-guided strategy and software, which Noble developed as a Ph.D. student, then pinpoint which electrodes can be turned off without loss of hearing fidelity—in fact, improving it. An audiologist uses this programming plan to create a revised custom program for that person's needs. The process is completely noninvasive—no surgery is required—and can be accomplished in one office appointment.

Under NIH funding, Noble plans to expand the technology in the future to test a wide range of position-dependent CI tuning techniques and to make these techniques available to the population of existing and new CI recipients.