Complexity of Registration Decisions during 3D Medical Volume Reconstructions.

Peter Bajcsy

Understanding Complex Systems Symposium, May 16-19, University of Illinois at Urbana-Champaign 2005.

We address the problem of optimal registration decisions during 3D medical volume reconstruction and their impact on (a) anticipated accuracy of aligned images, (b) uncertainty of obtained results, (c) repeatability of alignment, and (d) computational requirements. The registration decisions include (1) image spatial size (image sub-area or entire image), (2) transformation model (e.g., rigid, affine or elastic), (3) invariant feature (intensity, morphology or sequential combination of the two), (4) automation level (manual, semi-automated, or fully-automated), and (5) registration quality evaluation criteria (multiple metrics and methods for establishing ground truth). Our goal is to provide mechanisms for evaluating the tradeoffs of each registration decision in terms of the aforementioned impacts.

We present a medical registration methodology for making registration decisions that lead to registration results with well-understood accuracy, uncertainty, consistency and computational complexity characteristics. Furthermore, we have built software tools that enable making optimal registration decisions and we have provided access to them using web services. The tools are accessible using web services so that the registration decision support could be obtained by geographically distributed researchers while the computationally expensive operations would be performed at the location with supercomputing resources.