Alzheimer Staging and Detection User Interface using Neural Networks

Early detection and diagnosis are essential for controlling Alzheimer's disease because they enable effective therapies and long-term planning. Continuing endeavors are being made to gain deeper insights into the fundamental causes of the illness and to create improved therapeutic approaches. When the disease progresses, the person diagnosed would have trouble remembering recent events. At present, there is no known remedy for Alzheimer's disease. This work focuses on creating a user interface in the form of a mobile application for early detection and staging of Alzheimer's. This work contributes as an aid for both the diagnosed individual and the carer of the diagnosed. This work involves two phases in which the first phase involves the phycological questions and an audio recording which enhances the accuracy of the work. The detection phase employed a feedforward neural network with multiple inputs and a single output. This phase additionally requests the user under test to record their voice based on a reference audio. The correlation of the two audio files plays a significant role in the detection of the disease. The second phase is the staging in which the image of the brain scan is to be clicked and uploaded and the stage of the Alzheimer is detected if present. This staging phase of the disease is achieved by employing a convolutional neural network. This network was integrated to the user interface using a Tensor Flow Lite model, thus resulting in a high accuracy staging of the disease.