Performance Evaluation of Various Optimizers on Alzheimer’s Disease Classification Using Deep Neural Network

Authors

  • Sindhu T S

DOI:

https://doi.org/10.6977/IJoSI.202412_8(4).0002

Keywords:

Alexne, Googlenet, Squeezenet, Mobilenetv2, ADAM, RMSProp, SGDM, AD Classification

Abstract

The proposed work focuses on using transfer learning and CNN models for the classification of Alzheimer's Disease (AD) based on different classes of datasets. The goal is to improve the early diagnosis and classification of AD, which can contribute to better patient recovery and management. The study compares the performance of four different CNN models: AlexNet, GoogLeNet, SqueezeNet, and MobileNet V2. These models have been widely used in various computer vision tasks and have proven to be effective in image analysis. Additionally, three different optimizers are evaluated: Stochastic Gradient Descent with Momentum (SGDM), RMSProp, and ADAM. Optimizers play a crucial role in training deep neural networks, as they determine how the model updates its weights during the learning process. According to the results of the study, the MobileNet V2 model with the SGDM optimizer achieved the highest classification accuracy of 91% among all the tested classifiers. This suggests that this combination is particularly effective for AD diagnosis and classification based on the given datasets. The automated Alzheimer's disease classification system developed in this work has the potential to identify early signs and symptoms of the disease. Early detection is crucial because it allows medical professionals to intervene at an earlier stage, providing timely treatment and management strategies. By leveraging medical image analysis and transfer learning techniques, this system can contribute to more effective and efficient AD diagnosis, leading to improved patient outcomes.

Downloads

Published

2024-12-30