Ng, Chuen Rue and Than, Joel Chia Ming and Rijal, Omar Mohd and Kassim, Rosminah Md and Yunus, Ashari and Noor, Norliza Mohd (2019) Construction of 3D lung image morphology using 3D distance regularized level set. ARPN Journal of Engineering and Applied Sciences, 14 (1). pp. 109-113. ISSN 1819-6608,
Full text not available from this repository.Abstract
The World Health Organization (WHO) has stated in its report that lung disease is a wide spread in Malaysia which caused 2.77% of total death in Malaysia making it the 6th main cause of death in Malaysia. One of the lung diseases included in the list is interstitial lung disease (ILD). ILD includes an extensive group of disorders which leads breathing complications as a result of the alteration and fibrosis to anatomical structures in the alveolar structures. Therefore, diagnosis and analysis of ILD can be performed by segmenting the lung morphology on CT scans images. This study aims to construct a 3D lung image morphology using 3D distance regularized level set evolution (DRLSE). The 3D performance evaluations for normal lungs on average yielded better results than that of ILD lungs with a Dice's similarity coefficient of 93.19%. The constructed lungs from 3D DRSLE has good representation of the segmented normal lungs while suggesting deformities in segmented ILD lungs. © 2006-2019 Asian Research Publishing Network (ARPN).
| Item Type: | Article |
|---|---|
| Funders: | Universiti Teknologi Malaysia Research University, Ministry of Higher Education Malaysia |
| Uncontrolled Keywords: | 3D segmentation; CT scan; Level set |
| Subjects: | Q Science > QA Mathematics Q Science > QA Mathematics > QA75 Electronic computers. Computer science R Medicine T Technology > T Technology (General) |
| Divisions: | Faculty of Science > Institute of Mathematical Sciences |
| Depositing User: | Ms. Juhaida Abd Rahim |
| Date Deposited: | 20 Feb 2020 01:09 |
| Last Modified: | 20 Feb 2020 01:09 |
| URI: | http://eprints.um.edu.my/id/eprint/23852 |
Actions (login required)
 |
View Item |
