Multi-Class Diagnosis of Alzheimer's Disease Using Hybrid CNN-SVM Model

Recognizing the significance of early detection in Alzheimer's disease, a hybrid CNN-SVM model was developed to effectively identify and classify the condition. This model utilized a dataset of 3,000 MRI images, with an equal distribution of 1,000 images per class (non-demented, very mild demented, and demented). Through the application of preprocessing techniques to enhance and resize the images, with a focus on a specific region of interest, the model achieved exceptional performance. Notably, it attained an accuracy rate of 99.83%, sensitivity of 99.48%, specificity of 100%, and precision of 100%. These promising outcomes underscore the potential of machine learning techniques, exemplified by the hybrid CNN-SVM model, in significantly improving the early detection and diagnosis of Alzheimer's disease. Such advancements hold great promise for healthcare professionals and patients alike.