Applying Deep Learning Techniques for Alzheimer’s disease Classification: A Comparative study

Payyapilly Jonny, Jomol (2023) Applying Deep Learning Techniques for Alzheimer’s disease Classification: A Comparative study. Masters thesis, Dublin, National College of Ireland.

Preview	PDF (Master of Science) Download (1MB) | Preview
Preview	PDF (Configuration manual) Download (1MB) | Preview

Abstract

Alzheimer’s disease is a chronic brain condition that causes brain cell death, shrinkage, memory loss, and cognitive difficulties. The major cause of dementia is alzheimer’s disease. The functioning of the brain is affected by this disease. In the initial state of alzheimer’s disease the loss of memory is quite lower, and that in the late stage individuals receives cognitives disorders. This dangerous disorder requires to be handled in an early stage. Due to the increasing progress of this disease, it is crucial to diagnose the disease in early stage in order to limit further progression of the disease. Common machine learning algorithms do not improve the accuracy of predicting this disease, although deep learning techniques have increased the level of accuracy. My research on current research papers leads to the evidence for image processing, deep learning algorithms outperform common machine learning techniques. Image processing is performed on magnetic resonance imaging data to classify disease using deep learning algorithms. I propose a comparative analysis utilizing distinct deep learning techniques including CNN, RNN and transfer learning to classify alzheimer’s disease. Magnetic resonance data is employed for this research, were this study concentrates on individual performance of distinct deep learning methods. From the performance of distinct evaluations from algorithms on magnetic resonance data is easy to differ the algorithms performance and how they differ from each other. This paper highlights performance of distinct deep learning algorithms to classify stages of alzheimer’s disease. Convolutional neural networks(CNN), Transfer learning (TL) and Recurrent neural network(RNN) are the deep learning techniques which is used in this analysis. VGG19 architecture was utilized in CNN in order to perform AD classification. My transfer learning model underwent fine-tuning, which includes unfreezing the previously learned model and retraining it in order to enhance the model’s performance. The results acquired from classifying the images by using CNN model produced an accuracy of 0.93 percentage whereas transfer learning model used VGG16 architecture which showed 0.98 accuracy.

Item Type:	Thesis (Masters)
Supervisors:	Name Email Hafeez, Taimur UNSPECIFIED
Uncontrolled Keywords:	Image processing; Magnetic resonance images; deep learning; Machine learning; Alzheimer’s disease; dementia; Convolutional Neural Networks(CNN); Recurrent Neural Networks(RNN); Transfer learning
Subjects:	Q Science > QA Mathematics > Electronic computers. Computer science T Technology > T Technology (General) > Information Technology > Electronic computers. Computer science R Medicine > RC Internal medicine > RC0321 Neuroscience. Biological psychiatry. Neuropsychiatry Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Machine learning
Divisions:	School of Computing > Master of Science in Data Analytics
Depositing User:	Tamara Malone
Date Deposited:	23 May 2023 16:38
Last Modified:	23 May 2023 16:38
URI:	https://norma.ncirl.ie/id/eprint/6633

Actions (login required)

View Item