Quantitative Analysis of various Image Fusion techniques based on various metrics using different Multimodality Medical Images

Image Fusion is the process of combining two or more input images to obtain a resultant image which is rich in relevant information as compared to the original input image. The fusion technique finds its application in many areas: Robot Vision, Satellite Imaging, Medical Imaging, Remote Sensing and Defense imaging. In that Medical Imaging being the prominent ones. For efficient diseases detection and treatment, images from different modalities are combined using fusion techniques. This paper describes different techniques for fusion of multimodality images and the resultant images are analyzed using different quantitative measure. Initially, three different pairs of image are taken as input: Magnetic Resonance Imaging (MRI T2) and Computed Tomography (CT), Magnetic Resonance Imaging (MRI FLAIR) and Computed Tomography (CT), Magnetic Resonance Imaging (MRI T2) and Single-photon emission computed tomography (SPECT). Each pairs of images are fused together using fusion techniques namely Redundancy Discrete Wavelet Transform (RDWT), Mamdani type minimum-summean-of-maximum (MIM-SUM-MOM), Contourlet Transform (CONTRA) and Multiple Pulse Coupled Neural Network (MPCNN). The resultant is analyzed using quantitative metrics such as Entropy (EN), Standard Deviation (SD), and Mutual Information (MI). From the experimental results it is observed that MIM-SUM-MOM is efficient in providing better quality of images which is inferred from the values of EN. CONTRA gives better contrast as compare to other techniques which can be observed from thevalues of SD and also provides better retention of information from both the input images as displayed by the MI metric values.