Methods of Increasing Training Data for a 3D Neural Network for Alzheimer’s Disease Diagnosis

Alzheimer's disease (AD) is the most common cause of dementia among older adults. As it is associated with brain cell death and mass losses, deep learning and computer vision are suitable for magnetic resonance imaging (MRI) scans analysis for its detection. The study considers the 3D convolutional neural network to diagnose AD by MRI scans and the methods of increasing, augmenting and balancing the training set as the domain places high demands on it to build a correct diagnosing system. The methods discussed are datasets combining (for which ADNI and OASIS datasets are used) and spatial and intensity transforms of the training data (implemented in TorchIO). The purpose of the work is to estimate the impact of such modifications on the baseline model accuracy in AD diagnosis.