ABSTRACT
This paper focuses mainly on an efficient local stochastic search approach for the optimal registration transformation. A simultaneous perturbation stochastic approximation technique is successfully implemented on optimizing mutual information based similarity measures. The hill climbing search and the Nelder-Mead simplex direct search are also considered for the comparative purpose. Our registration experiments are associated with the pairs of optical sensor images, synthetic aperture radar images and medical multimodality images, which are misaligned by the rigid or affine transformations. The experimental results show that in general the local stochastic search effectively yields significant improvements on the optimal solution over the conventional search technique in terms of accuracy and robustness. The main contribution of this work is the first accomplishment of an efficient local stochastic search strategy on the mutual information based affine image registration scheme.
