Image registration algorithm based on generalized mutual information
Chengtao Feng · Transducer and Microsystem Technologies · 2008
Mutual information is widely used as a similarity metric for image registration.Usually similarity metric is based on Shannon definition of entropy.One generalized entropy named as Renyi entropy is used and a new image registration algorithm is proposed based on generalized mutual information.Firstly Renyi entropy with smaller q is used for global searches,here Renyi entropy can remove some unwanted local optimum;secondly,the local one is used to locate the global optimal solution by searching the current local optimal ones.The normalized mutual information measure based on Renyi entropy with q→1 is taken as the objective function.The experiment results show that the mutual information image registration algorithm based on Renyi entropy is better than the normalized mutual informatron image registration algorithm,and the speed of registration is improved.