Hybrid BAT-PSO optimization techniques for image registration

Medical images are progressively more being utilized as a part of healthcare and research. So, there is a wide interest for perfectly relating information in the distinctive images for diagnosis, treatment & basic science. This research paper proposed new registration techniques used to solve this problem, and describes the wide variety of applications to which these techniques are applied. Uses of image registration incorporate consolidating images of the same subject from diverse modal qualities, arranging temporal sequences of images to make up for motion of subject between scans, image guidance in the time of interventions & arranging images from numerous subjects in cohort studies. Images, which are correlated by a rigid body transformation, can be automatically registered by the existing registration algorithms. There has also been generous advancement in non-rigid registration algorithms, which can make up for tissue aligns images or deformation, from diverse subjects. These algorithms are not concentrating on transformation matrix. So, in this paper we concentrate on transformation angle information between registration images. Matching points was calculated between two images and transformation matrix. We use BAT-PSO algorithms to optimize Mutual information between two images. To make a comparison with the response of MI to translation & rotation the Multimodal medical images are utilized. The experimentations showed that FPMI is smoother & has less local fluctuations in comparing with MI. Registration results show that PSO-BAT do it better than existing methods to search optimal registration parameters.