Institutional Repository of Vilnius University
Kompiuterinė plaučių vėžio diagnostika /
Abstract
dc:descriptionThis study investigates the effect of malignant cancer and non-malignant cancer input data proportion on the results of computer aided lung cancer diagnosis systems (CAD), which use convolutional neural networks (CNN). This study overviews how CAD systems operate, what are the steps of CAD systems, their performance metrics, CNN usage in CAD systems and examples of such systems. Having not found data about TP/FP proportion effect on CAD systems results, it was decided to create an exemplary CAD system, which was trained 10 times with different TP/FP proportions. It was found that a big amount of TP input data (compared to the amount of FP input data) is harmful for the results of CAD systems and the ideal proportion for TP/FP input data is 1 to 5.
Degree
thesis:*- Grantor dc:publisher
- Institutional Repository of Vilnius University
- Year dc:date
- 2020
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Laucevičius, Marijus,
- 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:ELABAETD81706593&prefLang=en_US