An efficient registration method for medical images based on transform distance

Image registration is an excellent tool in different fields. Its medical applications are various in many clinical situations in order to analyze the patient's situation and to follow the localization of malignant sites. Generally, registration can be categorized according to several criterions. In this work, we are interested in the similarity function, for that we contributed a normalized dissimilarity index. This proposed approach is based on dissimilarity map that is a good tool to compare two images. We implemented this measure in a medical images example. The experimental results show that the proposed method is capable to align both binarized and gray-level images with higher precision.