Machine learning techniques for detection of glaucoma with optical coherence tomography angiography images
Glaucoma is one of the most common causes of irreversible blindness. It usually goes undetected until the disease progresses to an advanced stage. Therefore, this gives rise to a need for early detection of the disease to curb this issue before disease progression to an irreversible stage. Practit...
Saved in:
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Thesis-Master by Research |
| Language: | English |
| Published: |
Nanyang Technological University
2020
|
| Subjects: | |
| Online Access: | https://hdl.handle.net/10356/143892 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Nanyang Technological University |
| Language: | English |
| Summary: | Glaucoma is one of the most common causes of irreversible blindness. It usually goes undetected until the disease progresses to an advanced stage. Therefore, this gives rise to a need for early detection of the disease to curb this issue before disease progression to an irreversible stage.
Practitioners are required to analyze many data within fixed time frame. This is taxing and may result in a deviation of diagnosis over time due to fatigue. This report proposes to use machine learning techniques to detect glaucoma on OCTA images.
Two methods were proposed for the development of a CAD system. The first method used LPQ for feature extraction, feature selection using decision tree classifiers. Classification was done by using AdaBoost and achieved 94.3% accuracy.
The second method used the EQP for feature extraction, followed by using statistical t-tests for feature selection. Highest accuracy of 95.1% was achieved with an ensemble classifier. |
|---|