ABC Optimised CNN model for classifying Alzheimer's Disease stages

Alzheimer's disease is an advancing neurological illness that affects a significant number of individuals around the globe. Timely and precise identification of Alzheimer's disease is essential to ensure the efficacy of its treatment. Magnetic resonance imaging (MRI) can show structural alterations in the brain, it has become a viable technique for diagnosing Alzheimer’s Disease.However, manual MRI examination takes time and is prone to human error. To address these challenges, an algorithm for Alzheimer's Disease detection from brain MRI images using a convolutional neural network is proposed. The primary objective is to create a Convolutional Neural Network (CNN) and fine-tune the hyperparameters using the Artificial Bee Colony Optimization Algorithm. This is pursued in order to categorise the various stages of Alzheimer's disease through analysis of brain MRI images. The convolutional layers are in charge of collecting important characteristics from the images, and pooling layers to condense their spatial dimensions. Later, fully connected layers utilise these extracted features to classify the images accordingly. The goal is to reduce the loss experienced while learning which is a challenging step. When a model produces an output, it contrasts it with the intended output. It then calculates the disparity between the generated and intended outputs, aiming to minimise this difference and approximate the desired output more closely. This is accomplished by adjusting hyperparameters such as learning rate, number of convolution layers, batch size, dropout rate, number of epochs. To increase the model's accuracy, an optimization algorithm performs multiple iterations until a point of convergence. The approach is to develop a CNN model and apply hyperparameter tuning by using the Artificial Bee Colony optimization algorithm, to increase the efficiency of the CNN model. The model's accuracy was enhanced from 86.21% to 94% after applying the Artificial Bee Colony Optimisation algorithm.