Classification on Alzheimer’s Disease MRI Images with VGG-16 and VGG-19

Abstract

AbstractBalancing thoughts and memories of our life is indeed the most critical part of the human brain. Thus, its stability and sustenance are also important for smooth functioning. The changes in the structure can lead to disorders such as dementia and one such type of condition is known as Alzheimer’s disease. Multi modal neuroimaging like magnetic resonance imaging (MRI) and positron emission tomography (PET) is used for the early diagnosis of Alzheimer’s disease (AD) by providing complementary information. Different modalities like PET and MRI data were acquired from the same subject, there exists markable materiality between MRI and PET data. Mild cognitive impairment (MCI) is the initial stage with few symptoms of AD. To recognise the subjects which are capable of converting from MCI to AD is to be analysed for further treatments. In this research, specific convolutional neural networks (CNN) which are designed for classifications like VGG-16 and VGG-19 deep learning architectures were used to check the accuracy of cognitively normal (CN) versus MCI, CN versus AD and MCI to AD conversion using MRI data. The proposed research is analysed and tested using MRI data from Alzheimer’s disease neuroimaging initiative (ADNI).KeywordsAlzheimer’s disease (AD)Mild cognitive impairment (MCI)Cognitively normal (CN)Visual geometry group (VGG) and convolutional neural networks (CNN)