Medical Imaging Research Center
The role of imaging in healthcare is
continuously increasing. Recent innovations in medical imaging
technology have created a tsunami of imaging data, which is
revolutionizing diagnosis, therapy planning and follow-up, as well as
clinical, preclinical and biomedical research. Moreover, the rapid
adoption of digital image archiving and communication makes that large
image databases are readily available for multi-modal, multi-temporal,
and multi-subject assessment. A consequence is that accurate and
automated quantitative image computing has become indispensable.
The Medical Image Computing (MIC) research
group conducts application-driven research on quantitative image
computing, including image reconstruction, segmentation, registration
and visualization. Challenging applications are solved and validated in
a clinical environment in collaboration with clinicians and biologists.
These applications also serve as a basis to reveal the limits and
shortcomings of the state-of-the-art in medical image computing. Based
on this input, the research group gains a clearer insight into the field
and investigates novel problem-solving hypotheses.
Due to the complexity of medical imaging data
and the ambiguity inherent to limitations of the image acquisition
process, a prerequisite for quantitative image computing is the
availability of suitable parametric models that incorporate prior
knowledge about the typical appearance of the object of interest in the
image data. In medical applications these models need to be sufficiently
flexible to account for image appearance variations, such as normal
biological shape variability and pathological abnormalities. A powerful
strategy is to construct such models from the data itself by statistical
analysis of a representative training set of image instances. Current
flexible geometric models are represented by either a deformable shape
(e.g., set of landmarks, analytic curve, tetrahedral grid …) or a
deformable picture (e.g., atlas or gray value image itself). We propose
to combine both and further integrate them with other photometric and
physiometric information. Together with a suitable similarity measure
and optimization method this strategy offers a unified approach to
problems of image formation, image fusion and image quantification.
|
|
|
|
