Comparative Analysis of Deep Learning Techniques in Alzheimer's Disease Diagnosis: Trends, Challenges, and Future Directions.

Authors

Abstract

Alzheimer's disease is a slow, progressive neurological disorder that impacts the brain tissue and causes cells to die, the most common reason for dementia. It typically reduces brain volume, subsequently impairing several cognitive functions. We explored the phases of Alzheimer's disease diagnosis by thoroughly analysing the paper and highlighting the significance of pre-processing methods. Artificial intelligence progress has transformed medical image analysis by enabling automatic, precise identification of Alzheimer's disease through hybrid architectures, convolutional neural networks, and transfer-learning-based models. This article includes applications of Alzheimer's disease based on datasets, Artificial Intelligence models, limitations, pre-processing methods, challenges, and current research in this area. The article lists various medicinal modalities, risk factors for Alzheimer's disease, the disease's progression stages, and different criteria for evaluating the performance of Artificial Intelligence models. Accuracy, precision, confusion matrix, AUC, F1-score, ROC, recall, and MCC are some of the metrics covered in this paper. The paper discusses the number of comparisons of various deep learning techniques for Alzheimer's, new trends, limitations, recommendations, and future directions. Finally, the paper shows the ongoing practical outcomes and difficulties in Alzheimer's situations. Building on the analysis, this survey helps to cover different dimensions of Alzheimer's disease that have not been previously considered.

Tags

Topics