Alzheimer’s disease classification and detection by using AD-3D DCNN model

Deep learning techniques had achieved notability in the healthcare domain and are more specialized in medical imaging. Alzheimer’s disease (AD) enduring nervous system disorder affects elderly senior people with loss of cognitive processes and loss of memory. Early and précised detection of AD is essential for patient medical assistance and potential treatment. Since deep learning algorithms are adequate to analyze the enormous dataset and extract higher-level features from it, unlike traditional machine learning algorithms. This work presents a system-based deep convolutional neural network (DCNN) algorithm to detect AD and its stages. A DCNN and 3D densely connected convolutional neural networks (3D-DCNN) are used to perform the feature analysis and classification task. Finally, the features learned from the DCNN and 3D-DCNN are concatenated to classify disease. Alzheimer's disease neuroimaging initiative (ADNI) dataset is used for experimental analysis. The proposed AD-3DCNN model is compared to existing pre-trained models like Xception, inception V3, mobile Net, and dense Net and has recorded a highest accuracy for predicting different stages of AD. Apart from accuracy, various performance measures are used for evaluating system performance as precision, recall, and F1 score accuracy.