Classification of the Different Stages of Alzheimer’s Disease Using 3d Transfer Learning Networks

Abstract

Abstract Alzheimer's disease (AD) is a brain ailment that gradually impairs thinking and memory abilities as well as the capacity to do even the most basic tasks. A proper diagnosis of Alzheimer's disease (AD) is crucial for patient treatment, particularly in the early stages of the illness when patients can take precautions before suffering irreparable brain damage. In the proposed approach, the hippocampus area is identified as a biomarker by segmenting the region using 3D deep learning algorithms from pre-processed 3D MRI images. The classification of stages of AD was performed using 3D transfer learning techniques. The image quality parameters and classification parameters are derived from the resultant data for the analysis of the significant technique for the segmentation of biomarker, hippocampus and classification of stages of AD. The pre-processed 3D images are segmented with the 3D deep learning algorithms. Based on the image quality comparison the AD hybrid net design is determined to be more efficient for segmentation of hippocampus region. Later the segmented images are produced as input to the layers of different 3D transfer learning algorithms for classification of stages of AD. With the help of the Classification parameters, VGG Net-16 defined to be more appropriate for the process of segmentation. Thus, an efficient segmentation and classification technique for the identification of the different stages of AD is determined using image quality and classification parameters with high accuracy. These techniques are implemented to define a computer-aided diagnostic system for identification and prediction of AD.