Detailed project information

Image registration is an important task in medical image processing. It refers to the process of spatially aligning data sets, possibly from different modalities (e.g., magnetic resonance imaging (MRI) and computed tomography (CT)), different time points (e.g., patient follow-up), and/or different subjects (in case of population studies). It has many applications, and is frequently used in our institute as part of an image processing pipeline. In view of the enormous growth and advances in multi-modal imaging, automatic analysis of the sometimes large data sets becomes more and more important. Image registration, however, is a time-consuming procedure: currently the registration of a single pair of images requires 10-60 minutes, owing to the large number of degrees of freedom that need to be estimated.

The scientific goal of this application it two-fold: 1) investigate the use of supercomputing within the context of large patient population studies, i.e. requiring many registrations; 2) investigate the use of supercomputing for accelerating a single registration in the context of time-critical clinical procedures.