Alzheimer's disease diagnosis based on feature extraction using optimised crow search algorithm and deep learning

Alzheimer's Disease (AD) is long-term, progressive, degenerative cognitive illness and one of the causes of dementia. Dementia impairs an individual's ability to think, disrupting normal functioning. Conventional method of diagnosis is collecting symptoms from family members to analyse its impact and stage. MRIs are currently used worldwide for diagnosis and understanding how brain works. With recent advances in applying machine learning to medical images like MRI, it has become a key research discipline amongst experts and analysts. Existing methods of feature extraction from images include CNN, providing large number of feature sets that require great computation power and time to evaluate using traditional machine learning or deep learning algorithms. Consequently, we propose an Optimised Crow Search Algorithm (OCSA) for early detection of AD based on raw MRI image features, yielding a highly representative dense embedding. The mapping learned between this embedding and image labels resulted in diagnosing 98.62% accuracy.