Alzheimer Classification Based on Convolutional Neural Network and Vision Transformer

Alzheimer's disease (AD) is a progressive neurodegenerative disease with a hidden onset. Clinically, it is characterized by comprehensive dementia symptoms such as memory impairment, aphasia, loss of recognition, impairment of visuospatial skills, executive dysfunction. Clinical early diagnosis...

Full description

Saved in:
Bibliographic Details
Published in2023 International Conference on Image Processing, Computer Vision and Machine Learning (ICICML) pp. 329 - 334
Main Authors Mu, Xing, Zhang, Jiayi, Zhang, Kairui, Jin, Hao, Zhou, Xinglei, Xie, Zetao, Toe, Teoh Teik
Format Conference Proceeding
LanguageEnglish
Published IEEE 03.11.2023
Subjects
Alzheimer
Alzheimer's disease
CNN
Convolutional neural networks
Deep learning
Image processing
Magnetic resonance
Magnetic resonance imaging
Smote
Transformers
ViT
Online AccessGet full text

Cover

More Information
Summary:Alzheimer's disease (AD) is a progressive neurodegenerative disease with a hidden onset. Clinically, it is characterized by comprehensive dementia symptoms such as memory impairment, aphasia, loss of recognition, impairment of visuospatial skills, executive dysfunction. Clinical early diagnosis of Alzheimer's disease is not easy and precise. In this paper, we use the Convolutional Neural Network (CNN) and Vision Transformer (ViT) methods to learn and classify brain magnetic resonance images (MRI) of AD in the Kaggle public AD data set. We firstly use the Smote algorithm to solve the problem of uneven distribution of sample numbers, then we compare the accuracy of three models of CNN, ViT, CNN+ViT in the Alzheimer's disease data set. Finally, the highest accuracy rate for the test set is 97.43%. Such good result will play a certain role in assisting clinicians in making judgments.
DOI:10.1109/ICICML60161.2023.10424819