An improved 3D shape context based non-rigid registration method and its application to small animal skeletons registration.
Abstract
3D shape context is a method to define matching points between similar shapes. It can be used as a pre-processing step in a non-rigid registration. The main limitation of the method is point mismatching, which includes long geodesic distance mismatch causing wrong topological structure, and neighbors crossing mismatch between two adjacent points. In this paper, we propose a topological structure verification method to correct the long geodesic distance mismatch and a correspondence field smoothing method to correct the neighbors crossing mismatch. A robust 3D shape context model is generated and further combined with thin-plate spline model for non-rigid registration. The method was tested on phantoms and applied to rat hind limb and mouse hind limb skeletons registration from micro-CT images. Errors between the registered surfaces were reduced by using the proposed method. The robustness of the method is demonstrated.