Alzheimer's Disease Detection Using RNN and CNN Based Deep Learning Approach

Abstract

Alzheimer's disease is a very widespread form of dementia and is the fifth leading cause of death among those aged 65 and older. Furthermore, official data indicates a significant increase in the number of fatalities related to Alzheimer's disease. The early discovery of Alzheimer's disease might potentially enhance patient survival chances. Machine learning technologies used for magnetic resonance imaging have been effectively employed for the detection of Alzheimer's disease to accelerate the diagnosis process and support medical practitioners. Conventional machine learning methodologies need the use of specialized feature extraction techniques on MRI images, a procedure that is arduous and demands the involvement of a skilled practitioner. The use of deep learning as a self-sustaining feature extraction method might potentially automate the process and diminish the need for manual feature extraction. This paper proposes a pre-trained masked CNN deep learning approach as an independent feature extraction method for detecting Alzheimer's disease using magnetic resonance imaging (MRI) scans. Subsequently, a masked RCNN was evaluated against RCNN and Faster RCNN using several criteria, including accuracy, to determine which performed superiorly. The results indicated that the proposed model surpassed existing state-of-the-art methods by achieving a higher degree of accuracy. The suggested methodology attained an accuracy of 98% when applied to algorithms using the MRI ADNI dataset.