AD-OCT: An Early Alzheimer Disease Detection Method via OCT Images using Advanced Feature Extraction and Clustering Techniques

for efficient management and treatment whereas the researchers focus on the various approaches including biomarkers, imaging and techniques like Magnetic resonance imaging (MRI), Optical Coherence Tomography (OCT) and advanced Deep learning (DL) algorithms to identify the early signs of the disease before significant cognitive decline occurs. DL holds noteworthy promise in proceeding Alzheimer's research and patient care. Conventional models have traditionally faced challenges in capturing the intricate details and subtle features within OCT images that are crucial for early detection of AD, leading to superior performance. Imbalanced datasets, where positive cases of Alzheimer's are far fewer than negatives, often bias models toward the majority class, which hampers accurate detection. Moreover, current models often struggle to accurately pinpoint pathological structures within OCT images, which are essential for dependable diagnosis, presenting a significant hurdle and furthermore, limiting the generalization of models to unseen data and affecting their clinical utility. By overcoming these issues, Authors have approached an Advanced Feature Extraction and Integration Techniques by this method Authors can perceive the AD earlier. Initially Authors provide the OCT images in the Integrated Residual Enhanced Attention Module (IREAM) which combines the strengths of residual learning and attention mechanisms and the extracted features are now refined and clustered using the Cluster Feature refinement (CFR) and Clustered Feature Fusion (C2F) which improves the quality and the extracted features are merged to get a unified set. Finally, Author’s model is compared with the existing model and Authors calculate the balance accuracy, Recall, AUC, Dice coefficient and F1-Score of the proposed model.