Interpolation of 3D Medical Images in Direction of Slice Thickness

抄録

The equal spacing in 3 dimensions (3D) is needed in many applications of medical images, which were usually processed by interpolating from the sequence of cross-sectional slices. Typically, the spacing between slices is much greater than the resolution on the slice. In order to get the equal spacing, people usually use the direct linear interpolation method based on point matching. However, the direct linear interpolation method makes the images blurred on their edges. In this paper, a new interpolation method based on optimization search for corresponding points was developed. We built up the corresponding relation between the interpolated images by optimization process, which was based on the image features with the gray value, the gradient of the gray value and the gradient direction in neighboring slices to get the paired points. Then gray value of the interpolated point in new slice was determined by linear interpolation from the paired points. The simulation results showed that our method was more accuracy than the direct interpolation methods.