A NEW GLOBAL REGISTRATION APPROACH OF MEDICAL IMAGING USING VECTOR MAPS

In this paper, we introduce a novel global image registration approach using vector distance functions (VDF´s). Edges of the source and target objects are used to represent their images. The VDF´s of these edges are calculated as an implicit representation of these boundaries. An energy is formulated to measure the differences of these vectors. A variational formulation is considered to estimate the transformation parameters. Promising results in 2D and 3D are demonstrated to show the efficiency of the proposed algorithm.