Wavelet based signal processing techniques for medical image fusion

Abstract

Recently signal and image processing have been central to researchers and scholars through present various applications and solve many problems in different fields in our life. This thesis presents signal processing algorithm for multi-modal medical images by fusion technique. Medical image fusion has been used to derive texture from multi-modal medical image data. The idea is to improve the image content by fusing images like computer tomography (CT) and magnetic resonance imaging (MRI) images. This derived texture can be assisted by medical examiner for various purposes such as, diagnosing diseases, detecting the tumor, surgery treatment, and clinical treatment planning system. Our object to get more as possible better image fused high quality and clearer. Previous fusion based on the spatial domain and another depends on the frequency domain, both these strategies have disadvantages like contrast reduction, weak quality, artifact, and ringing. Therefore researchers in medical fusion field attempt to solve these problems by many algorithms are presented and are competed to improve previous results. Hence, this work present an algorithm based on Discrete Wavelet Transform (DWT) to obtain the scale and detail coefficients of the various images. Different fusion methods are also used comparing ; Non-linear fusion rule (NLFR), average mean value (AMV), maximum absolute rule (MAR), and Weighted Condition Value (WCV) to correlate the coefficients each method is used separately then produce the last result by Inverse Discrete Wavelet Transform (IDWT) which based on single level transform. The novelty in this thesis are using two strategies, first one, deal with match measures are calculated as a whole to select the wavelet coefficients coming from different wavelet transform filters banking ,Second once using NLFR method, output results to compare with the chosen method so as to determine which is better. The medical fusion system implemented by MATLAB software, and analyzed the results done by Petrovic Fusion Algorithm (PFA). The method yields high scores the conventional methods. Overall this method has high potential for a better application of fusion in the medical imaging field.