NeuroSpectra: An Innovative Multi-Class Alzheimer's Disease Classification with ConvNeXt Transfer Learning Model

The disease known as Alzheimer's is a disorder of the nervous system that can be quite devastating. It is characterized by the gradual degeneration of brain cells, leading to the impairment of memory, cognitive function, and the ability to perform daily tasks. While a cure for Alzheimer's remains unclear the timely identification of the disease is of utmost importance in effectively managing its symptoms. Machine learning, specifically deep convolutional neural network methods, enhanced by transfer learning strategies, have emerged as powerful tools for tackling the complex issues associated with analyzing brain imaging data, particularly concerning Alzheimer's disease. The main aim of this categorization study is to encourage innovation and technology for long-term growth in the healthcare sector. This paper introduces a Convolutional Neural Network model built upon the ConvNeXt architecture. This proposed model can classify the brain images into four classes. Alzheimer's dataset has been utilized for conducting this brain image categorization task. This dataset includes 6400 images belonging to four classes demented, mild, very mild and non demented. The model has demonstrated a high level of accuracy, with a rate of 98.89% in classifying the images. The achieved values for the model's precision, recall, and F1-score are 99.07%, 98.99%, and 0.99, respectively. This research can be utilized in the medical field. By enhancing the early identification and monitoring of this disease, the research advances healthcare access.