Medical Image Fusion of Multi Modal Images using Random Block Selection Method

摘要

Image fusion is a process of obtaining a singleimage by combining multiple images. Since, the sourceimages are obtained from various detectors at dissimilartimes, some information will be missing in one sourceimage and that missing information may present inanother source image. Hence, the physicians are in theneed of combining the fine information from both sourceimages into a single one. The objective of image fusion isproviding utmost information that are missing in thesource images. In this study CT (Computed Tomography)and PET (Positron Emission Tomography) images are getfused. The CT imaging provides anatomic informationwhereas the PET image provides functional informationof the body. The proposed method fused the images usingcontourlet based random block selection method withMAX fusion rule. It is proved that the proposed methodprovides better result with less number of computations.Experimental results are taken by writing MATLAB code.To prove the quality of fused image, performancemeasures such as Peak Signal to Noise Ratio (PSNR),Structural Similarity Index Measure (SSIM) and MutualInformation (MI) are used.