Multimodal Medical Image Elastic Registration Using Mean Shift

Abstract

Estimating landmarks corresponding plays a key role in landmark-based multimodal image registration. In this paper, a novel landmarks corresponding estimation in multimodal image registration using mean shift algorithm is proposed. Edge feature potential is defined to transform images from intensity feature space to edge structure feature space. Image corner points are detected as candidate landmarks. Mean shift iterations are adopted to search the most probable corresponding point positions in the two images based on the edge structure feature. Moreover, mutual information between two local regions is computed to eliminate mis-matching landmarks. Finally, the source images are transformed by compact support thin-plate spline interpolation. Experiments show that the precision in location of corresponding landmarks is satisfied. The proposed technique is feasible and rapid shown in the experiments of various multi-modal medical images registration.