Medical Image Fusion Using Otsu’s Cluster Based Thresholding Relation

Abstract

Medical imaging is the method of making pictorial illustrations of the inner parts of a human body. Multimodality medical images are required to help more precise clinical information for specialists to make do with medical diagnosis, for example, Computer Tomography (CT), Magnetic Resonance Imaging (MRI), and X-ray images. The fused images can regularly prompt extra clinical data not evident in the different pictures. Another preferred standpoint is that it can diminish the capacity cost by putting away simply the single combined image rather than multisource images. The necessities of MIF are that the resultant fused image ought to pass on more data than the distinct images and should not present any artifacts or mutilation. In this paper, the existing DTCWT-RSOFM method is modified by using the Otsu’s cluster based thresholding relation with fuzzy rules are carried out on each sub band independently and merges the approximation and detailed coefficients. The inverse DTCWT is used to obtain the fused image. The results of simulation indicate that our method has attained maximum PSNR of 54.58 whereas the existing one with maximum PSNR of 34.71. Thus, there is an improvement of 35% in PSNR. It’s quite evident from visual quality of the proposed method output that, the edge based similarity measures are preserved fine. The simulation results indicate that modified technique has attained maximum EBSM of 0.9234 whereas the existing one with maximum EBSM of 0.3879. Thus there is an improvement of 42% in EBSM. The fused image offers improved diagnosis without artifacts.