Abstract—In this study, liver Magnetic Resonance (MR) images and Computerized Tomography (CT) images were combined with wavelet-based image fusion methods to facilitate expert decision-making in identifying the type of liver focal lesions. For this purpose, 46 MR and 46 CT images were used belong to 36 different patients. These images include different type of focal liver lesions samples that cysts, Hepatocellular Carcinoma (HCC), Colagiocellular Carcinoma (CCC), Focal Nodular Hyperplasia (FNH), liver metastases and hemangioma. For the fusion, three different fusion rules including average rule, maximum selection rule and multiplication rule was applied to images and the results were compared. When the results were visually examined, it was observed that the multiplication rule was more successful. In addition to the visual results, the performances of fusion results are compared using, Peak to Noise Ratio (PSNR), Accuracy (ACC), Entropy (EN) and Fusion Factor (FF) metrics.

Index Terms—image fusion, liver, MR imaging, CT imaging, discrete wavelet transform, image fusion rules

Cite: Saim Ervural and Murat Ceylan, "A Comparison of Various Fusion Methods for CT and MR Liver Images," Journal of Image and Graphics, Vol. 5, No. 2, pp. 59-63, December 2017. doi: 10.18178/joig.5.2.59-63