Segmentation-based interpolation of 3D medical images

Abstract

This paper introduces a new interpolation algorithm based on images segmentation. Firstly, the algorithm obtains the regions of air, soft tissue and skeleton through segmenting images. Secondly, the algorithm uses matching interpolation in the same density regions, and scales the size of region as the interpolation data to interpolate image in the different density regions. The new image basically satisfies the requirements of medical image interpolation. Compared with linear interpolation, the new algorithm greatly improves the quality of image. The interpolation can be effectively used to construct 3D volume models. segmentation-based interpolation; 3D image; threshold segmentation; matching point pair ©Springer-Verlag 2004.