Tools
Tools
Mastamardi, Prajwal Shivalingappa (2021) Anomaly identification in chest radiography. Masters thesis, Dublin, National College of Ireland.
Preview |
PDF (Master of Science)
Download (878kB) | Preview |
Preview |
PDF (Configuration manual)
Download (916kB) | Preview |
Abstract
COVID-19 is a recently identified coronavirus that causes mild to moderate respiratory sickness in many people. However, the new disease has a significant impact on the elderly and those with underlying health problems. To understand the process of evolution of the model, several outcomes from studies on the issue were studied. Computer vision can perform tasks like object recognition and picture classification, which may be used to classify and identify lung images. The goal of this study is to employ deep learning techniques to categorise and detect COVID-19-affected medical lung pictures. This work suggests using deep learning methods such as VGG-16 and ResNet- 50 in identifying and classifying the anomalies present in medical lung images. Various classes of data were modelled from binary classification with anomaly dataset, three label classification and four label classification. Models trained on balanced class of anomaly dataset yielded highest result on common test data, VGG-16 model had an accuracy of 81% with recall value of 91%, ResNet-50 had an accuracy of 86% with recall of 81 for covid cases.
| Item Type: | Thesis (Masters) |
|---|---|
| Subjects: | Q Science > QA Mathematics > Electronic computers. Computer science T Technology > T Technology (General) > Information Technology > Electronic computers. Computer science R Medicine > RA Public aspects of medicine T Technology > Biomedical engineering Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Machine learning |
| Divisions: | School of Computing > Master of Science in Data Analytics |
| Depositing User: | Tamara Malone |
| Date Deposited: | 22 Feb 2023 17:38 |
| Last Modified: | 02 Mar 2023 09:26 |
| URI: | https://norma.ncirl.ie/id/eprint/6221 |
Actions (login required)
 |
View Item |