Institutional Repository of Vilnius University
Plaučių ligų diagnozavimas rentgeno nuotraukose /
Abstract
dc:descriptionIn this study methods of automatic pneumonia detection in chest X-rays (CXR) that are found in literature are presented. These methods are based on traditional image processing techniques and convolutional neural networks. One of presented is the method proposed by A. Sharma and his colleagues. It is based on traditional image processing and Otsu thresholding. We proposed new algorithm that was based on A. Sharma’s one. In this algorithm we used HOG, Haar-like and LPB approaches to detect and cut lungs from whole image as well as Otsu thresholding. By this algorithm we created application. While testing it with real CXRs we noticed that this method could help recognize pneumonia, because for healthy lungs it showed less corrosion in lung masks and they were smoother than for lungs with present pneumonia. However, sometimes it was not able to correctly recognize lungs and cut CXRs with present pneumonia. Also, sometimes even in healthy lungs Otsu thresholding was showing some disintegration that are common for pneumonia affected lungs.
Degree
thesis:*- Grantor dc:publisher
- Institutional Repository of Vilnius University
- Year dc:date
- 2020
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Vaišys, Laurynas,
- Contributors dc:contributor
-
- Plankis, Tomas
Rights
dc:rights- Statement dc:rights
-
- info:eu-repo/semantics/openAccess
- Language dc:language
- lit
Identifiers
dc:identifier.*- Repository record dc:identifier
- https://repository.vu.lt/VU:ELABAETD81705861&prefLang=en_US