Shape-based interpolation of porous and tortuous binary objects

Abstract

This paper presents a feature-guided approach for shape-based interpolation of porous and tortuous binary objects. The feature points derived from the boundaries of the candidate source objects are matched non-linearly. The intermediate objects are obtained by appropriately blending the warped source objects. A robust outlier-rejecting, non-linear point matching algorithm based on thin-plate splines is used for establishing the feature correspondence. The proposed scheme correctly handles objects with holes, large offsets and drastic invaginations. Preliminary results suggest that this approach could be used to significantly enhance the sparse Talairach-Tournoux brain atlas.