The main task of medical 3D reconstruction is to build accurately an isotropic volume consisting of cubic voxels from 2D serial cross sections. However, the spacing between the slices is typically much greater than the size of a pixel within the slice. Therefore, interpolation between the slices is of great importance. This paper presents a new kind of contour interpolation method in order to overcome the drawbacks of the existing methods. It assumes being applied after a segmentation step extracting the contours, and after a contour association step in order to avoid pathological cases. Moreover, it is based on the geodesic metric instead of Euclidean metric, and allows the interpolation of complex contours with concavities.
展开▼
