An efficient and accurate interpolation strategy for multi-dimensional functions

We propose an alternative approach for performing multi-dimensional interpolation of bandlimited functions. Exploiting the properties of the sampling patterns of the underlying functions of interest, our approach accomplishes the task of multi-dimensional interpolation by use of the fast Fourier transform (FFT) and lower-dimensional interpolation. The proposed technique is easy to implement and more accurate than the most commonly used multi-dimensional linear interpolation. When applied to two- and three-dimensional (2D and 3D) problems that arise in medical imaging, the proposed approach is faster and more accurate than are the conventional bilinear and 3D-linear interpolation approaches.