Image registration is one of the important issues in image processing which has been of much interest in last decades. Its significant application is in medical imaging where it has been used from 1980-1990. Image registration is the alignment of two or more different images of a view or an object, such that their corresponding pixels coincide with each other. These images might be acquired by several sensors or by a single sensor in different viewpoints, moments, or other different conditions. The goal of image registration is to find a transform function by which the floating image is aligned to the reference image. Every image registration algorithm require three basic ingredients: 1) a spatial transformation model which determines the set of possible solutions, 2) an objective similarity measure which estimates the quality of each potential solution and 3) an optimization algorithm which looks for the best solution. To find the desired transformation function and to determine how similar the two images are, a similarity measure must be used. The parameters of the transformation function are obtained by optimizing the value of the similarity measure between two images. In other words, the similarity measure takes maximum value when the optimum transformation function is applied to the floating image. Mutual information is the most popular similarity measure in medical image registration. It is one of the basic concepts of information theory which indicates the dependency of two random variables (or two images). The most important and critical step in calculating the mutual information of two random variables (or two images) is to find their joint probability distribution. Since the images are samples of two-dimensional continuous signals, calculation of their joint probability distribution is not error-free. Therefore, selection of a reliable method for estimating the joint probability distribution is of great importance. Usually, an interpolation method is employed for calculating the joint probability distribution of two images. Using such interpolation algorithms may decrease the accuracy of registration. In this thesis the reasons behind popular usage of mutual information as a similarity measure is extensively investigated. In other words, we intend to answer the following question: Why mutual information is such popular in registration, while numerous distances have been defined in the literature? Also, the effect of some interpolation methods in the accuracy of registration is studied. A new approach is introduced to decrease such undesirable effects. Keywords: Image registration, Floating image, Refrence image, Similarity measure, Mutual information, Interpolation