Paper Key : IRJ************045
Author: Aleena Thomas
Date Published: 14 Oct 2023
Abstract
Early detection of Alzheimers disease is important as the brain is the main affected organ in (Alzheimers disease)AD. Early detection helps in minimizing the effect of AD. MRI (Magnetic Resonance Imaging) is the main method used in diagnosing AD. However, the detection of AD with this process is time-consuming. With Rapid advancement in AI technology, early detection of AD is possible using various techniques. This paper mainly focuses on detecting AD using various types of Convolutional Neural Network(CNN) models. A comparative study is done using an MRI image dataset with various CNN models to obtain the best accuracy. The dataset used in this paper is MRI images collected from the Alzheimers Disease Neuroimaging Initiative (ADNI). This dataset was divided into 3 classes: CN, MCI, and AD. Preprocessing of the dataset was conducted to increase the effectiveness of the dataset and to obtain more accuracy. The study was done on 7 various CNN models, including LeNet, AlexNet, Inception_V3, ResNet-50, VGG16, VGG19 and DenseNet-121.The models were evaluated based on accuracy, precision, and F1 score. From the comparison study, it is observed that Dense Net achieved the highest accuracy of 96.01%, and the lowest accuracy which is 85.07% was achieved by Inception_V3.
DOI LINK : 10.56726/IRJMETS45247 https://www.doi.org/10.56726/IRJMETS45247