Multimodal image registration by minimizing Minkowski distance
Wufan Chen · Journal of Circuits and Systems · 2008
Based on the work of R. Gan et al 2004 and the theory of Minkowski inequality, Minkowski distance (MD) between two joint intensity distributions as a new concept of distance measure is proposed. MD is used to scale the similarity of two images in the multimodal image registration. With the introduction of MD, the convexity and smoothness of the objective function in the proposed method are much better than the ones of R. Gan's work. And the proposed method can efficiently solve the local extreme problem and increase the registration success rate remarkably. Many experimental results demonstrate that the accuracy of the proposed method is comparable with the mutual information maximization technique based image registration.