Enhanced independent component analysis using symmetric orthogonalization for alzheimer’s disease image classification
Also available in printed version
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Master's thesis |
| Language: | English |
| Published: |
Universiti Teknologi Malaysia
2025
|
| Subjects: | |
| Online Access: | https://utmik.utm.my/handle/123456789/50899 |
| Abstract | Abstract here |
| _version_ | 1854975069123510272 |
|---|---|
| author | Gopinathaan Rajamohan |
| author2 | Mohd. Saberi Mohamad, supervisor |
| author_facet | Mohd. Saberi Mohamad, supervisor Gopinathaan Rajamohan |
| author_sort | Gopinathaan Rajamohan |
| description | Also available in printed version |
| format | Master's thesis |
| id | utm-123456789-50899 |
| institution | Universiti Teknologi Malaysia |
| language | English |
| publishDate | 2025 |
| publisher | Universiti Teknologi Malaysia |
| record_format | dspace |
| record_pdf | Abstract |
| spelling | utm-123456789-508992025-08-21T00:12:29Z Enhanced independent component analysis using symmetric orthogonalization for alzheimer’s disease image classification Gopinathaan Rajamohan Mohd. Saberi Mohamad, supervisor Computing Also available in printed version There is an unmet medical need for identifying neuroimaging biomarkers for early and accurate diagnosis for management of Alzheimer's disease (AD). In AD studies, magnetic resonance imaging (MRI) has shown its powerful ability to distinguish AD, mild cognitive impairment (MCI) subjects from and healthy controls (HC). FastICA is an independent component analysis (ICA) based algorithm that has some limitations in terms of choosing a suitable initial separating matrix due to deflationary orthogonalization process that can lead to extracting insignificant features. In order to overcome these challenges, a FastICA algorithm using symmetric orthogonalization to extract features from segmented MRI images is proposed in this research. First, all MRI scans are aligned, normalized and tissue segmented by statistical parametric mapping (SPM). Then, FIRST and FREESURFER tools were used for further subcortical segmentation of amygdala and hippocampus region on grey matter images. Following this, our enhanced FastICA algorithm was implemented to these images for extracting features that were fed into a support vector machine that discriminates AD, MCI and control subjects. This study comprised of 416 and 509 subjects from open access series of imaging studies (OASIS) and Alzheimer's disease neuroimaging initiative (ADNI) depositories respectively. Classification result was validated using several training and testing ratios using cross validation. 90-10 ratio obtained the best classification result on both databases. In ADNI, classification of HC and MCI yielded the best accuracy result of 88.34% and for OASIS, 91.44% attained for classification of AD and HC. The experimental outcomes demonstrated that the proposed method achieved better results compared with existing methods in classifying AD, MCI and HC subjects fahmimoksen UTM 138 p. Thesis (Sarjana Falsafah, Bidang Penyelidikan : Sains Komputer) - Universiti Teknologi, 2017 2025-03-14T03:40:28Z 2025-03-14T03:40:28Z 2017 Master's thesis https://utmik.utm.my/handle/123456789/50899 vital:111281 valet-20180603-09086 ENG Closed Access UTM Complete Unpublished Completion application/pdf Universiti Teknologi Malaysia |
| spellingShingle | Computing Gopinathaan Rajamohan Enhanced independent component analysis using symmetric orthogonalization for alzheimer’s disease image classification |
| thesis_level | Master |
| title | Enhanced independent component analysis using symmetric orthogonalization for alzheimer’s disease image classification |
| title_full | Enhanced independent component analysis using symmetric orthogonalization for alzheimer’s disease image classification |
| title_fullStr | Enhanced independent component analysis using symmetric orthogonalization for alzheimer’s disease image classification |
| title_full_unstemmed | Enhanced independent component analysis using symmetric orthogonalization for alzheimer’s disease image classification |
| title_short | Enhanced independent component analysis using symmetric orthogonalization for alzheimer’s disease image classification |
| title_sort | enhanced independent component analysis using symmetric orthogonalization for alzheimer s disease image classification |
| topic | Computing |
| url | https://utmik.utm.my/handle/123456789/50899 |
| work_keys_str_mv | AT gopinathaanrajamohan enhancedindependentcomponentanalysisusingsymmetricorthogonalizationforalzheimersdiseaseimageclassification |