Alzheimer’s Detection And Categorization using a Deep-Learning Approach

Alzheimer's disease (AZM) is a deadly neurology syndrome caused by memory loss. It is the most common kind of dementia. It is a syndrome that begins with modest memory loss and gradually deteriorates your thinking abilities, occasions, and ability to execute day-to-day routine tasks. It affects the areas of the brain that govern cognition, memory, and language. It has the potential to have a significant influence on a person's ability to carry out everyday responsibility. Because no drug or cure has yet been discovered, it is critical to forecast it at an early stage. The prediction of Alzheimer's disease is done manually by evaluating MRI scans; nevertheless, the risks of human mistake are quite high with this method. The purpose of this research is to improve prediction accuracy in order to detect AZM at an early stage and prevent the patient from progressing to a severe state of the disease. In this research, a Convolutional Neural Network (CNN) with Softmax classifier (i.e., CNN+Softmax) has been used to identify and categories the AZM MRI into four distinct classes, namely mild demented (MD), moderate demented (MOD), non-demented (ND), very mild demented (VMD). The proposed CNN+Softmax model is then compared by replacing the dense layer of CNN to Support Vector Machine (SVM) classifier, i.e., CNN+SVM and the work is calculated on the various grids which includes accuracy, precision, recall and f1-score. On test data, the presented approach surmounted the CNN+SVM model with a classification accuracy of 98.59 percent.