Full metadata record
| DC Field | Value | Language |
|---|---|---|
| dc.contributor | Department of Computing | en_US |
| dc.contributor.advisor | Chung, Fu-lai (COMP) | en_US |
| dc.creator | Guo, Chen | - |
| dc.identifier.uri | https://theses.lib.polyu.edu.hk/handle/200/11334 | - |
| dc.language | English | en_US |
| dc.publisher | Hong Kong Polytechnic University | en_US |
| dc.rights | All rights reserved | en_US |
| dc.title | On the use of 3D convolutional neural network for Alzheimer's disease diagnosis | en_US |
| dcterms.abstract | Alzheimer's disease is a common chronic neurodegenerative disease. The progression of Alzheimer's disease develops slowly and become worse over time, and the patients will gradually lose the ability to live independently. Therefore, early diagnosis and interventional treatment of Alzheimer's disease can significantly delay the progression of cognitive decline and improve the living quality of patients. Currently, the computer-aided Alzheimer's disease diagnosis is emerging as a fast and efficient approach. However, training a robust neural network from scratch requires a large number of ground truth data, which is almost impossible to achieve in Alzheimer's disease diagnosis fields. Therefore, in this study, an adaptation method to improve classification accuracy on smaller datasets were tested. For this purpose, a 3D convolutional neural network was trained to classify Alzheimer's disease and normal patient using public MRI dataset. The pretrained 3D convolutional neural network achieved classification accuracy of 0.74 and 0.62 in source and target domain datasets, respectively, by using fine-tuning strategy. Then, the classification accuracy in target domain was improved from 0.62 to 0.71 by using our adaptation method, which iteratively fine-tune the model using unlabeled data. In real world scenario, medical image datasets usually consist of a small amount of ground truth data and the remaining unlabeled data. Our experimental results indicated that fine-tuning with a limited unlabeled dataset on a pre-trained model through proposed adaptation method can significantly improve classification accuracy. | en_US |
| dcterms.extent | v, 41 pages : color illustrations | en_US |
| dcterms.isPartOf | PolyU Electronic Theses | en_US |
| dcterms.issued | 2020 | en_US |
| dcterms.educationalLevel | M.Sc. | en_US |
| dcterms.educationalLevel | All Master | en_US |
| dcterms.LCSH | Alzheimer's disease -- Diagnosis | en_US |
| dcterms.LCSH | Machine learning | en_US |
| dcterms.LCSH | Medicine -- Data processing | en_US |
| dcterms.LCSH | Hong Kong Polytechnic University -- Dissertations | en_US |
| dcterms.accessRights | restricted access | en_US |
Files in This Item:
| File | Description | Size | Format | |
|---|---|---|---|---|
| 5805.pdf | For All Users (off-campus access for PolyU Staff & Students only) | 1.41 MB | Adobe PDF | View/Open |
Copyright Undertaking
As a bona fide Library user, I declare that:
- I will abide by the rules and legal ordinances governing copyright regarding the use of the Database.
- I will use the Database for the purpose of my research or private study only and not for circulation or further reproduction or any other purpose.
- I agree to indemnify and hold the University harmless from and against any loss, damage, cost, liability or expenses arising from copyright infringement or unauthorized usage.
By downloading any item(s) listed above, you acknowledge that you have read and understood the copyright undertaking as stated above, and agree to be bound by all of its terms.
Please use this identifier to cite or link to this item:
https://theses.lib.polyu.edu.hk/handle/200/11334