3D Medical image surface rendering using an optimal interpolating function

This paper reports on the use of kriging, a geostatistical interpolation process, for computing 3D medical image surface renderings. Kriging is known to be the “best” linear unbiased estimator. We briefly describe the kriging process and its use of regionalized variables to model zones of influence in spatial data. We used kriging in combination with the “marching cubes” algorithm to successfully compute surface renderings of medical imaging modality data. Our results indicate that kriging is a better interpolating function than trilinear or tricubic interpolation. We conclude with suggestions for further research in this area.