A new MRI and PET image fusion algorithm based on BEMD and IHS methods

Considering the kind of information contained in brain images, they can be divided into two main groups, the first group are those which exhibit structure of brain providing high spatial resolution like magnetic resonance imaging (MRI) images, and the second group are the images which indicate functionality of brain with low resolution like positron emission tomography (PET) images. Image fusion is of great importance in combining the strengths of these groups to provide high quality images containing maximum possible of information. In this paper, we propose a new method based on bidimensional empirical mode decomposition (BEMD). We use this method to decompose the MRI and also the intensity component of PET image after it is decomposed with IHS method. In the next step, we compose useful details of both images and discard leftover. Studying properties of the output and input images using visual and numerical evaluation methods, confirms the effectiveness of our method. To show the strength of our proposed method we also compared some of its properties such as standard deviation (STD), entropy, spectral discrepancy (SD), average pixel intensity (API), average gradient (AG) and overall performance (O.P) with PCA and wavelet fusion methods. Our method has better results than both of the others.