Abstract
dc:descriptionFire detection in video stream has been researched for couple of decades already. Artificial neural networks open new possibilities to improve upon previously achieved results in fire detection. There already exists convolutional neural networks that try to solve this problem, but this research investigates a new not yet highly explored type of artificial neural network: capsule networks and their usage for fire detection. During this research a capsule neural network was created named FireCapsules that detects fire and it was compared to newest convolutional neural networks designed for fire detection. After this research we can conclude that capsule networks can be used for fire detection and that they are not worse than their convolutional neural network counterparts. Capsule networks are not yet highly explored topic so further research is a must in order to achieve better results.
Degree
thesis:*- Grantor dc:publisher
- Institutional Repository of Vilnius University
- Year dc:date
- 2020
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Bedulskis, Augustas,
- Contributors dc:contributor
-
- Valaitis, Vytautas
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:ELABAETD81705749&prefLang=en_US