FUSION OF MULTIMODAL ABDOMINAL CANCEROUS IMAGES AND CLASSIFICATION USING SUPPORT VECTOR MACHINE

In medical field, the modality based image analysis is attaining much importance due to the clinical data has to be processed to analyze various outcomes. Fusion of multimodal images is performed to combine all relevant information from single or multiple imaging modalities into a new single. In this work fusion of different modality images like Computed Tomography (CT), Magnetic Resonance Imaging (MRI), Positron Emission Tomography (PET) of abdomen cancer is carried using Laplacian Pyramid fusion rule and Multi-resolution Singular Value Decomposition (MSVD) fusion rule. The fusion performance is analyzed using various quality metrics like Entropy, Fusion Factor (FF), Standard Deviation (SD), Structural Similarity Index and Correlation measure. Fused images are classified as benign or malignant lesion using Support Vector Machine (SVM) classifier.