A novel implementation of medical image registration and fusion using ASM-SSIF

In medical imaging a single image of a targeted organ is captured using different imaging modalities such as MR (Magnetic Resonance) and CT (Computed Tomography), to get desired information from the region of interest. For better diagnosis the information obtained from these two modalities has to be combined into a single image. Image fusion is a process of integrating useful or complementary information from multiple images into a single image. In this work, we proposed a novel method for image fusion with biomedical images i.e., MR and CT images, by formulating a convex optimization problem. The optimization problem uses an Active Slope Meagerness (ASM) regularizer with statistics based steered image filtration (SSIF). Precise registration is needed by the MR and CT image fusion, while the misalignment is quite hard to obviate during the preprocessing step. To surmount this, we had proposed a robust and new approach for both registration as well as the fusion of MR and CT images. Our approach focuses on simultaneous registration during the fusion procedure. Initially, the MR image is focalized to enhanced resolution, which permits us for more precise registration of images. At the same time, fusion of images can be done precisely by extinguishing the misalignment gradually. We iteratively persist these two procedures until the convergence. The performance of the algorithm is judged both qualitatively and quantitatively. The Proposed ASM-SSIF method is compared with the existing methods. The simulation results showed that our algorithm has given greater performance by enhancing the overall fusion quality of MR and CT images in terms of image quality assessment. Specifically, our proposed approach is shown to be much more powerful on the medical data sets of real-world with pre-registration errors.