Comparative Analysis of Image Registration Using Pixel, Wavelet and Translation Method
Mashiur Rahman, A. H. M. Zadidul · Electrical and Electronic Engineering · 2015
Image Registration is the process of estimating an optimal transformation between two images. There are many application of image registration now a day such as in FMRI motion correction, geometric distortion in EPI, alignment of images obtained at different time or with imaging parameters, formation of composite functional maps, mapping of PET/SPECT to MRI image. In this work the image registration analysis is done using the affine transformation model in our proposed three method image registration using ‘pixel’, ‘wavelet’ and ‘translation’ method. A comparative experiment in between these methods based on the analysis time and root mean square deviation. If the image is displaced from its place which means it can be rotated, displaced from its original axis vertically or horizontally this method can detect the displacement and can get the original image. At first we do a random rotation and then we find an image which is larger in size from the original image. Then we again rotate the image by reshaping the size of the image. The registered is finding from again rotating the image in a loop and tries to find a relation between them. From this relation we get the value at which the image was rotated or displaced from the original image. Then we reverse the process and get the final result.