Abstract
dc:descriptionIn this work, I focus on the generation of Lithuanian font, having only several different same font letters. In the literature analysis part, I revise the supervised, unsupervised, and reinforcement learning. In the next section Generative Adversarial Networks and related topics such as different GAN models. stabilization, expansion, evaluation of results, and image difference evaluation are analyzed. During the experiment, the MC-GAN model is improved to generate not only Latin but also Lithuanian letters. The experiment is run on two machines: first being MacOS with Intel i5 7500 processor. Second is a virtual Linux machine from Google Cloud using video card Tesla T4. Conclusions are that improved MC-GAN can be used for Lithuanian font generation and the video card is much faster for data training.
Degree
thesis:*- Grantor dc:publisher
- Institutional Repository of Vilnius University
- Year dc:date
- 2020
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Naruševičius, Lukas,
- Contributors dc:contributor
-
- Būtėnas, Linas
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:ELABAETD81705372&prefLang=en_US