3-D Interpolation of Medical Images Based on the Optimal Cubic Convolution Kernel

Abstract

The cubic convolution interpolation is usually employed to 3D interpolation of medical images. In the paper, the interpolation principle of cubic convolution is analyzed systematically, and essential relationship among the different cubic convolution interpolation methods is formulated. Based on these, a 3D interpolation method of medical images that is based on the optimal parameter of sharp control (PSC) is presented. The approach makes full use of the local characteristic of medical volume data, and the local optimal PSC is obtained by the iterative computation, and then the volume data is interpolated using the cubic convolution under this parameter in one time. The experimental results show the method presented in the paper not only can improve the interpolation accuracy effectively, but also is robust.