Application of Convolution Neural Network to Detect the Stages of Alzheimer Disease for Magnetic Resonance Imaging

The major form of dementia that cause memory loss at the final stage of the disease is known as Alzheimer. The autonomous techniques of artificial intelligence are used to detect the Alzheimer disease at the earlier stage. The major focus of the work is to find an automatic method in order to find and classify the different stages of Alzheimer disease. In this work a deep neural network g model is used to detect the Alzheimer diseases and non-Alzheimer diseases. The input data is taken from the open source ADNI dataset. Initially all the input images are pre-processed. The total input images are segregated into training and testing dataset. The convolution neural network (CNN) is used to detect and classify the stages of Alzheimer such as demented, non-demented, mild dementia and very mild dementia. The method is evaluated based on the metrics of precision, accuracy and recall. The quantitative results are compared with other methods that have performed classification of Alzheimer disease. The overall detection accuracy of 97.8% is achieved in the proposed work.