Research Of Medical Image Registration Based On Mutual Information

Medical image registration is one of the important parts on medical image processing, provide doctors to utilize the information of different model medical images from the same patient, so the accuracy of medical diagnosis was improved greatly. It not only can be used for diagnosis and treatment , can also be used for tracking and evaluating the effect of treatment. Now, many registration methods which have been brought forward can be divided into two kinds: the registration methods base on image features and the registration methods based on gray statistics. The first kind of registration method is simple and can be arrived easily, but the registration precision is limit to the segmentation when the image features are acquired. Mutual information registration model based on gray statistics, and the mutual-information based registration method is of high accuracy and robustness without the need for preprocessing of images, Hence it represents the trend of registration.Firstly, it is described in the paper that the background, development actuality and practices of medical image registration. Then the paper describes the principle of medical image registration. And the detailed steps how to finish the registration also are analyzed, in which the geometry transform, the interpolation method, the similarity measure, the optimized algorithm and their effects on registration are thoroughly discussed. Meanwhile, the classification of the registration method and the concerned evaluation are surveyed. Next, the paper studies the medical image registration based on mutual information, the background and the foundation concept of it are introduced, and the difficulties of this method is carried on the discussion.Medical images registrations usually use the local optimizing algorithms. But these local algorithms are easy to fall into the local supreme value, so it will lead to the wrong registration. The paper uses genetic algorithm and Particle Swarm Optimization algorithm, both of which have good global searching ability. Because of the earliness, slow searching speed and long running time, genetic algorithm has some defects when it is applied in medical images registration. The paper improves the genetic algorithm from coding, genetic operating ,and so on. It improves the capability of the algorithm. From being combine with Particle Swarm Optimization algorithm, the paper design a hybrid of medical image registration algorithm, improving the registration process. Finally, the registration experiments of mutual information using the improved optimized algorithm are done under Matlab7.0 environment, in which the medical images of the human brain are taken as the testing data. The simulation achieves the medical image registration based on mutual information, which results confirm the accuracy and the robust of the registration algorithm, then some advantage and shortcoming existing in the improved optimized algorithm is summarized by the compare with other optimized algorithm.