Alzheimer’s Disease Computer-Aided Diagnosis: Histogram-Based Analysis of Regional MRI Volumes for Feature Selection and Classification

摘要: This paper proposes a novel fully automatic computer-aided diagnosis (CAD) system for the early detection of Alzheimer's disease (AD) based on supervised machine learning methods. The novelty approach, which is histogram analysis, twofold: 1) feature extraction process that aims to detect differences in brain regions interest (ROIs) relevant recognition subjects with AD and 2) an original greedy algorithm predicts severity effects these regions. takes account progressive nature affects structure different levels severity, i.e., loss gray matter found first memory-related areas such as hippocampus. Moreover, proposed generates reduced set attributes allows use general-purpose classification algorithms. In particular, approach assesses ROI image separability between classes order identify ones greater discriminant power. These will have highest influence decision at final stage. Several experiments were carried out segmented magnetic resonance images from Disease Neuroimaging Initiative (ADNI) show benefits overall method. CAD achieved competitive results highly efficient straightforward way.