Multimodal Brain Images Fusion using Cultural Algorithm Optimized Multispectral Features

Abstract

Medical images can be acquired through different techniques (modalities), which have their own application areas; some of them provide information on the functional activity, while others contain only anatomic information. Usually, in the first case, images have low spatial resolution while in the second case have a higher resolution. However, the analysis of medical images often requires the evaluation of more than one modality; in order provide the specialist with more information for decision making as well as for the analysis and the treatment of diseases. Image fusion aims to combine information from the same sensor or different sensors, so that the image fused retain the information content of each individual image. In remote perception, when multispectral images are analyzed, it is very important to preserve the content of spectral information of each of the bands. The challenge is to obtain good quality images that allow us to extract as much amount of information possible, for which it is sometimes necessary to enhance or modify the image to improve its appearance or combine images or portions thereof to combine the information. An ideal fusion of multispectral images and the band panchromatic will result in a new series of bands with greater spatial resolution and equal spectral content. This paper proposes a PCA, DWT and cultural optimized entropy based DWT fusion with the evaluation parameters; arithmetic mean (SM), Maximum value ( ) and Minimum value ( ).