A New Approach for Alzheimer's Disease Diagnosis by using Association Rule over PET Images

Abstract

disease is usually diagnosed from patient history and clinical information. Finding appropriate technologies and early detection of AD is of fundamental importance for early treatments. A set of PET images is selected for the study. In order to ensure that a given voxels in different images are refer to the same position the images are normalized using Spatial Normalization which are subjected to noise filter using Butter worth Filter. Intensity Normalization is required to perform direct image comparisons in which the intensity is normalized to an Imax value. Based on Activation Estimation the Region of Interest (ROI) is achieved which are subjected to Association Rule Mining by specifying the minimum support and the confidence values. Finally Computer Aided Diagnosis (CAD) method performs the image classification with verified rules based on threshold. The comparison of previous methods is performed the early finding of AD.