Applying Deep Learning Methods to the Ai Expert System on Alzheimer`s Disease for the Elderly

Abstract

With the advent of an aging society, the number of people who are being afflicted with Alzheimer's disease is also on a gradual rise. However, an effective medical treatment to contain the contraction of the disease still does not exist. There is a widespread lack of understanding about dementia among the general populace, and the amount of available research into early pathological prediction of Alzheimer's disease is also quite low. This study proposes a testing system to detect cases of dementia, which is designed to assist doctors in diagnosing the disease. The system can be applied to embedded devices and mobile devices in the hospital in the future to promote the development of artificial intelligence in the medical field and improve diagnosis efficiency. On the basis of minuscule size and a small number of parameters, lightweight convolutional neural networks can be deployed on devices with finite memory and computing without connecting any cloud platform to avoid the breach of image data and ensure the quality of its security. For this purpose, three common lightweight convolutional neural networks are used in this study -MobileNet V2, NASNetMobile, and ShuffleNet V2. In cases where the parameters are identical to those corresponding to other conditions, the Alzheimer's disease predictive identification is applied to use open-source magnetic resonance imaging (MRI) scans obtained from the Kaggle platform. The results of the study indicate that MobileNet V2 exhibits the highest prediction accuracy (80.78%). Additionally, the system proposed in this study can be integrated into physicians' workflows during the diagnosis of Alzheimer's disease, thereby making their medical judgments more accurate. It can, therefore, address the tedious and time-consuming nature of the current methods of diagnosis of the disease, improve the efficiency of the medical treatment of patients, and improve the possibility of early detection of the disease.