BICIR: Boundary-Constrained Inverse Consistent Image Registration Using WEB-Splines

Abstract

We present a new registration method called boundaryconstrained inverse consistent image registration (BICIR). This method is novel because it performs boundary constrained intensity based image registration by combining surface- and intensity-based registration. The method registers a region of interest and ignores everything outside the region eliminating interference caused by surrounding regions. The boundaries of the two objects are first registered using the transitive inverse-consistent manifold registration (TICMR) described in [4], This provides the boundary conditions, which are used to compute the displacement over the object using the Element Free Galerkin Method (EFGM). The EFGM solution is used as an initialization and is fine-tuned using the intensity information inside the object. The transformations are represented with shape specific basis functions that have local support. These basis functions are defined only over the region of interest and help improve the computational performance. The BICIR method was tested for a number of 2-dimensional test images. The method was validated using the similarity cost, relative overlap and inverse consistency error. © 2006 IEEE.