Statistically optimal interslice value interpolation in 3D medical imaging: theory and implementation
Rob W. Parrott, Martin R. Stytz, Phil Amburn, D. Robinson · 2003
Describes a technique for statistically optimal interslice interpolation of scalar values for use in three-dimensional medical image rendering. The interpolation technique is based upon kriging, which is known to be the best linear unbiased estimation technique for spatially distributed data. The authors present the results obtained using kriging in the object space preprocessing operation of slice interpolation by slice-value interpolation. As a byproduct of the technique, kriging calculates the estimation error for the interslice values. This makes it possible to quantify the interpolation error in slices computed by the estimation technique.>