An Age Impairment Ad Dementia on Various Stages Depicting the Severity on Patient Using a Deep Learning Model

Abstract

In Medical field there are major affected parts in brain such as cancer in brain, disorders, mental pressure and whack on brain and so on. The main part of brain malformations we name it as tumors that are disorders that may leads to dementia’s and Alzheimer disease's that mainly identified on old people which has no treatment and 99% leads to death as results. These disorders are mild degenerative that happened in brain nervous which can be of various parts such as any layer of neural networks. In general the brain consists of receiving the commands given to brain and the moment reacts to the command which consists of multiple hidden neuron that are produced as result to output layer, where the multilayer perceptron have a major role to work on the process of overtime of deteriorate that happened in change of people in their behavior, character even the infections that leads to death. These issues on brain can be treated to predict in various techniques such as computer vision, artificial intelligence, Internet of things, Machine learning and deep learning. Our research focuses on deep learning techniques on age impairment AD dementia to find the various stages of affected parts such as Mild dementia, Moderate dementia and severe dementia that causes to death. The main techniques that are used to predict the analysis of aging people in better results to prioritize the severity is principal component analysis (PCA). There are datasets such as ADNI and non ADNI, oasis which has produced its results of identifying the patients affected part on Mild cognitive impairment (MCI) stage that will be forgetting the memories of activities such as cycling, walking, remembering the known things etc. In our proposed method we are using a best fitting model for reducing the squared vertical similarities also reduced from the brain's image from MRI which is classified to identify the orthogonal forms to calculate the matrix point of the affected parts of brain image and form the correlation as comparison of training data with the Eigen value that are formulated by covariance using the decomposed value of matrix point of the value that gives the exact prediction as the component score. The exact assumptions of analysis can avoid the over fitting problem and transform the value to test the best result using the PCA on the convolutional neural network model for classifications.