Comparison of CNN Models with Different Plane Images and Their Combinations for Classification of Alzheimer’s Disease Using PET Images

Abstract

Alzheimer’s disease (AD) is an irreversible progressive cerebral disease, with most of its symptoms appearing after 60 years of age; to date, it is not possible to completely cure AD. Positron emission tomography (PET) is a functional molecular imaging modality and has been shown to be a powerful tool to understand AD-related brain changes compared to magnetic resonance imaging (MRI) and computed tomography (CT). In recent years, various machine learning methods, including deep learning, methods have been proposed for the classification of AD using PET brain images. However, most of these methods only use features of the axial plane, disregarding features of the sagittal and coronal planes despite the brain being a three-dimensional structure. This is because which plane image provides more useful information for the classification of AD remains unclear. In this paper, we compared the classification results of convolutional neural network (CNN) models with different plane images and their combinations are compared. Experimental results show that the CNN model with the coronal plane image as the input image achieves the best results (about 89%).