Medical Images Separation and Fusion Based on Artificial Neural Network

Abstract

During the last few decades, the field of medical image processing has been closely related to neural network methodologies and their applications. In the present investigation a 512×512 Computed Tomography (CT) and Magnetic Resonance Imaging (MRI) images for different region of the brain are registered to eliminate the dimensionality differences between the two images, then separated both of them by fast-fixed point algorithm after truncation of each image in to almost 1000 image patches of 15×15 dimension and transform them to 1-D and order them into row-wise fashion as well as reducing the entered data of lesser interest by Principle component analysis (PCA), finally applying the fusion process using different methods. The result shown that the differently defined brain regions can be separated using batch approaches for both CT and MRI and could be a powerful and accurate diagnostic tool, especially, for surgical and radiotherapy, planning and oncology treatment after a suitable fusion process is carried out on it.