A new algorithm for interpolation between grayscale serial slice images, such as from CT, is presented. The algorithm extends shape-based (SB) binary image interpolation to shape-based interpolation of grayscale images (SBIG). Unlike algorithms, such as linear (L) or cubic spline (CS) interpolation, which rely only on pixel position, SBIG makes essential use of object distance and morphology to interpolate between pixels and structures of similar shape and intensity which may differ in size and position from slice to slice. For reasonably low noise MRI, CT, and Cine CT gray-scale images, results are superior visually and quantitatively (15%) to interpolation based solely on (x,y) proximity, particularly as the interslice spacing is increased. More importantly, while both L and CS interpolation demonstrate characteristic low-pass smearing of object edges and detail, these features are preserved and well approximated with SBIG. As a result, reconstructed coronal and sagittal slices from a densely interpolated image volume using SBIG demonstrate significantly clearer representation of anatomical structures and less RstaircasingS than those created using either L or CS interpolation. Clipping artifacts due to nonoverlapping structures or rapid changes in image brightness are minimized using simulated three-dimensional distance maps.