A new image processing algorithm for computer aided prediction of glaucoma in Ghana
| dc.contributor.author | Adjei, Prince Ebenezer | |
| dc.date.accessioned | 2016-11-03T16:00:29Z | |
| dc.date.accessioned | 2023-04-19T12:59:57Z | |
| dc.date.available | 2016-11-03T16:00:29Z | |
| dc.date.available | 2023-04-19T12:59:57Z | |
| dc.date.issued | MAY 2016 | |
| dc.description | A thesis submitted to the Department of Computer Engineering, Kwame Nkrumah University of Science and Technology in partial fulfillment of the requirements for the degree of Master of Philosophy Department of Computer Engineering. | en_US |
| dc.description.abstract | Glaucoma is a term that describes a family of eye diseases that damage the optic nerve. About 8% of the entire population of Ghanaians above the age of 30 suffer from what is known as Primary Open Angle Glaucoma (POAG). The nemesis of POAG is blindness. POAG usually presents no symptoms, thus the only way to prevent blindness from POAG is early detection and consistent monitoring of progression of the disease, as well as the effect of medications. Diagnosing POAG in Ghana is typically done by eye specialists using a method known as ophthalmoscopy, which involves examining the interior of the eye (the fundus) with a magnifying instrument. The challenge with this however is with the availability of experienced eye specialists, especially in rural areas where POAG is most prevalent. This thesis employs the use of a relatively new type of Artificial Neural Network known as the Pulsed Coupled Neural Networks together with thresholding techniques and a simple multi-layer back propagation neural network to predict glaucoma from fundus images. We propose a new algorithm for extracting key glaucoma prediction parameters – the vertical cup to disc ratio and the inferior superior nasal temporal rule assessment. The system then subsequently predicts glaucoma. A total of 166 fundus images from Soyuz Medical Imaging and Diagnostic Center in Accra, Ghana was analyzed by the department of optometry and visual science in KNUST by primary eye care specialists, and set as the standard for measuring the reliability of the algorithm. The system correctly extracted and measured the vertical cup-to-disk ratio with a root mean square error of 0.11. Further, it achieved a specificity of 91.4% and a sensitivity of 88.6%. The system is thus reliable, and can be employed in a mass screening application. | en_US |
| dc.description.sponsorship | KNUST | en_US |
| dc.identifier.uri | https://ir.knust.edu.gh/handle/123456789/9579 | |
| dc.language.iso | en | en_US |
| dc.title | A new image processing algorithm for computer aided prediction of glaucoma in Ghana | en_US |
| dc.type | Thesis | en_US |
Files
Original bundle
1 - 1 of 1
Loading...
- Name:
- Prince Ebenezer Adjei .pdf
- Size:
- 2.95 MB
- Format:
- Adobe Portable Document Format
- Description:
- Full Thesis