Institutional Repository of Vilnius University
Lietuvių gestų kalbos atpažinimas iš vazdo įrašo /
Abstract
dc:descriptionThis paper deals with the task of recognising Lithuanian sign language from video data. A website was created to collect a dataset of Lithuanian sign language. The dataset consists of 10 classes, 25 different individuals and 2500 videos of 60 frames each. Next, two architectures were trained using the developed dataset: a hybrid model of convolutional neural networks coupled with a long short‑term memory model and a model of video vision transformers. The trained hybrid convolutional neural network model coupled with the long short‑term memory model achieved a maximum accuracy of 58.87%, while the trained video vision transformers model achieved a maximum accuracy of 80%.
Degree
thesis:*- Grantor dc:publisher
- Institutional Repository of Vilnius University
- Year dc:date
- 2024
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Kondrotas, 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:ELABAETD210645883&prefLang=en_US