International Journal on Recent and Innovation Trends in Computing and Communication ISSN: 2321-8169 Volume: 3 Issue: 2 500 - 503 _______________________________________________________________________________________________ 500 IJRITCC | February 2015, Available @ http://www.ijritcc.org _______________________________________________________________________________________ Comparison of Different Image Fusion Techniques for 2D MRI Images Prof. Mrs.Megha Sunil Borse 1 1 Dept. of Electronics &Telecommunication Cummins College of Engineering for Women Karvenagar, Pune, India 1 megha.borse@cumminscollege.in Prof. Dr. Mrs. Shubhangi B. Patil 2 2 Dept. of Electronics Dr. J.J. Magdum College of Engineering Jaisinghpur,Sangli, India. 2 sbp_jjm2004@yahoo.co.in Abstract—Image fusion is the process of combining relevant information from two or more images into a single image. The resulting image contains more information than the input images. Thus data fusion combines partial and varied information which is present in multiple images and forms a single image having the collective features of all the input images. It has two main aims which are removal of partial redundant data, as all sources provide information about the same phenomenon ;and Other is the complementarities between data as each source provides a different view about the same phenomenon. Two techniques are implemented for image fusion which are Wavelet Transform and Fuzzy Logic. The results of these techniques are compared based on Entropy, Standard Deviation and Mutual Information. Keywords -Fuzzy Inference System (FIS); Membership functions(mf); Entropy (EN); Standard Deviation(SD); Mutual Information (MI) __________________________________________________*****_________________________________________________ I. INTRODUCTION The fusion techniques find many applications in real life. As single fused image is stored instead of multiple images the storage space required is reduced. Fast image retrieval is possible, since knowledge base has less number of stored images (fused). It also reduces the possibility of data replication, i.e .the patient data stored is reduced. Image fusion technique has become a common application used within medical diagnostics and treatment. Fused images may be created from multiple images which are obtained from the same imaging modality or by combining information from multiple modalities, such as magnetic resonance image (MRI), computed tomography (CT), positron emission tomography (PET), and single photon emission computed tomography (SPECT)[8]. In radiology and radiation oncology, these images are used for different purposes[10]. For example, CT images provide more information related to bony structure while MRI images are typically give more information about the tissues hence used to diagnose brain tumors. Thus image Fusion improves the quality of information from a set of images. Important applications of the fusion of images is in the area of medical imaging, microscopic imaging, remote sensing, computer vision, and robotics. Quality of this fused image we can judge in terms of performance parameter like Entropy (EN), Standard Deviation (SD), and Mutual Information (MI). Recently, Discrete Wavelet Transform (DWT) and Fuzzy Logic Based Image Fusion techniques have been popular fusion of images. These methods give much better result than simple averaging, maximum, minimum. This paper is organized into five sections. In section 1 Different Image Fusion techniques are mentioned. In section 2 Image fusion with wavelet transform using average and maximum coefficient method is explained. In section 3 Image fusion technique using Fuzzy Logic is explained. Section 4 consists of comparison of these fusion techniques. Section 5 comprises of conclusion and Results. II. IMAGE FUSION TECHNIQUES Image fusion has important applications in many different image processing fields such as satellite imaging, remote sensing and medical imaging. Image fusion method can be broadly classified into two groups[5] – A. Spatial domain fusion method B. Transform domain fusion A. Spatial Domain Spatial domain methods does direct processing on the pixels of an input image. A mathematical expression for spatial domain processing is given by the equation1 g(x ,y) = T[f(x ,y)] ..........(1) Here ,the original image is given by f(x, y),and processed image by g(x, y) and T is an operator which is applied over neighbourhood of (x, y). The neighbourhood is defined about a point (x, y) and it is of a square or rectangular type with sub image area centred at (x, y). Spatial domain filtering can be applied for smoothing and sharpening purposes. Here Fuzzy Logic based method is implemented which combines the pixel intensities of the source images depending on the Fuzzy rule selected. This fusion method gives the better result with preservation of data from input or source images. B. Frequency Domain Drawback of Spatial distortion can be minimized by frequency domain[9] method of image fusion. The multi resolution analysis is a useful tool for analysis of remote sensing images. The discrete wavelet transform is useful technique for fusion.