Institutional Repository of Vilnius University
Vaizdų antraščių kūrimas naudojant mašininio mokymosi metodus /
Abstract
dc:descriptionWe encounter visual materials every day and in large quantities, for example– in commerce and marketing. Many of people also encounter them in their work environment, such as radiologists or social media content creators. An automated image captioning tool could ease and expedite people’s daily routines. The goal of this work is to apply machine learning methods used for image captioning tasks to evaluate and compare the results obtained and to perform image captioning in Lithuanian. The research employs the architecture of regional convolutional and bidirectional recurrent networks. Experiments were conducted with well-known datasets: Flickr8k, Flickr30k and MS COCO, with unseen images, and with the Flickr8k dataset translated into Lithuanian. The evaluation of the generated captions used the BLEU, METEOR, ROUGE-1, and ROUGE-L metrics.
Degree
thesis:*- Grantor dc:publisher
- Institutional Repository of Vilnius University
- Year dc:date
- 2024
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Bagdonaitė, Rugilė,
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:ELABAETD210643214&prefLang=en_US