International Journal of Computer Theory and Engineering, Vol. 1, No.2,June2009 1793-8201 - 173 - Abstract— We investigated the registration of medical images based on the Normalized Tsallis entropy using mutual information measure. A prerequisite for successful registration is unambiguous maximum of mutual information. We discuss the framework of our algorithm with Normalized Tsallis entropy as the core. Further we propose a type II fuzzy based technique to select the optimal Tsallis parameter q which provides the best alignment. Consequently, specific instances of image registration involving rigid affine transformations were studied. Registration was applied to clinically acquired mammogram. The accuracy was compared with several other techniques. Our algorithm shows promising results. Further, the Need for Pre-registration in mammogram is discussed in detail. Our algorithm can be effective enough to replace Shannon and Tsallis entropy based affine registration. Index Terms— Tsallis entropy, Shannon entropy,Normalized Tsallis entropy, Joint intensity distribution,image registration, Powell optimization, Mammogram. I. INTRODUCTION Image registration is the determination of geometrical transformation that aligns points of one view of an object with the corresponding points in another view of that object. i.e. output is a geometrical transformation which is simply a mathematical mapping from one points to points in second. There are many image registration methods and they can be classified into many ways. Mutual information (MI)based technique is the most popular technique, because MI does not rely on the intensity values directly to measure correspondence between different images, but on their relative occurrence in each of the images separately and co-occurrence in both images combined. MI is insensitive to one-to-one intensity transformations and is capable of managing positive and negative intensity correlations simultaneously. MI is not based on intensity differences or intensity correlation, like other pixel-based registration criteria. The purpose of the registration (rigid) using MI is to reduce global spatial differences between corresponding images caused by the positioning difference during their J. Mohanalin is with the Indian Institute of Technology , Kanpur, India (corresponding author phone number: 0988914581). P.K.Kalra, is with with the Indian Institute of Technology , Kanpur, India He is now with the Department of Electrical, He was Head of Department of Electrical. Nirmal Kumar is with the Indian Institute of Technology , Kanpur, India. He is the Chief Doctor of the Hospital inside IIT kanpur. acquisition. By finding a set of parameters, tx, ty, and θ, capable of maximizing the MI between the two images, the best registration location is found. It has been found that MI based similarity measure can still fail for certain clinical images. Improved performance is detected by various normalization schemes. Current Validation also supports this theory. This paper shows the application of normalized Tsallis entropy (NTE) as a new method of rigid registration instead of traditional Shannon entropy (SE). The Paper is organized as follows. Section 2 Deals with the literature review of recent work in rigid image registration. Section 3 explains basics of our algorithm, and details of Normalized entropy. A new method to find optimal Tsallis parameter using Type II fuzzy set is proposed. The proposed algorithm is checked for mammograms with known simulated deformation and results with several validation techniques were shown in section 5. Section 6 contains the conclusion. II. BRIEF SURVEY OF EXISTING WORKS IN IMAGE REGISTRATION: The Shannon-MI has received tremendous attention [1] and is robust and accurate in registering images. Viola et al. [2,3] was the first one to propose Image registration based on MI. Studholme et al. [4] introduced normalized mutual information (NMI) to rigidly register multi-modal images with different fields of view; Pluim et al. [5] used a gradient-based term along with the MI to avoid the problems of local maxima. Rueckert et al. Maintz et al. [6] used correlation for multi-modal image registration. In recent years NMI has proven to be a robust and accurate similarity measure for image registration [7]. In [8], Rangarajan et al. proposed feature point registration with MI. Rényi entropy [9, 10] and Tsallis mutual information (Tsallis-MI) [11,12,13], are the other two entropy techniques showing promising results, consequently their properties make them conductive to medical image registration. MI, was commonly used to solve global spatial differences between mammograms [14]. III. REGISTRATION ALGORITHM Let one of the images selected be the reference image, R(x,y) and other image be target image T(x,y). The MI based methods state that, for two images that are to be registered, the value of their MI will be maximal if the images are geometrically aligned. The NMI of two images is expressed in terms of the entropy of the images. Entropy is a measure of Mutual Information based Rigid Medical Image registration using Normalized Tsallis entropy and Type II fuzzy index. Mohanalin, Prem Kumar Kalra and Nirmal Kumar