Brain tissue volume estimation to detect Alzheimer's disease in magnetic resonance images

Volume estimation of brain tissues such as the White Matter, Gray Matter and Cerebrospinal Fluid is an important task in brain image analysis and also used to diagnose neurological and psychiatric disorders. In this work, brain tissue volume reduction is estimated to detect Alzheimer's disease (AD) using magnetic resonance images. The proposed method initially applies Hue Saturation Value-Based Histogram Thresholding Technique to segment the brain tissue. After that, brain volume is estimated using the pixel counting-based method (PCBM) to detect AD. The proposed method was investigated with images obtained from T1-weighted images of cognitive normal (CN) /normal (N) and AD images from Minimum Interval Resonance Imaging in Alzheimer's Disease and Alzheimer's Disease Neuroimaging Initiative and T1- and T2-weighted real-time images collected from a medical diagnostic clinical imaging center. The estimated brain tissue volume between the AD and CN/N brain tissue clearly quantifies the brain tissue reduction and it is compared with existing automatic estimation method statistical parametric mapping (SPM). Comparing to SPM, our PCBM method accurately estimates the brain tissue volumes and can be used as a potential tool to detect AD using MR imaging data.