Efficient Visualization of Fiber Tracking Uncertainty based on Complex Gaussian Noise
Klein J, Hahn H, Erhard P, Peitgen H, Althaus M, Rexilius J, Leibfritz D
MeVis - Center for Medical Diagnostic Systems and Visualization GmbH
Visualizing the uncertainty of fiber tracking is an important new challenge in the area of DTI. We present a new method that allows for an efficient computation of diffusion weighted images with user-defined noise, which are used to analyze the tracking uncertainty resulting from image noise. In contrast to the bootstrap method, our technique needs only a single data set so that acquisition time and the time for computing the artificial data can be reduced dramatically. Our visualization of the resulting fiber sets as well as our measurements show that bootstrap noise can be simulated appropriately by complex Gaussian noise.