Institutional Repository of Vilnius University
Teksto ir dokumentų struktūros atpažinimas naudojant grafų konvoliucinius tinklus /
Abstract
dc:descriptionAn important area in document structure recognition – recognition of text paragraphs. Past research based on graph convolutional networks addresses this problem, but is limited to the detection of physical text paragraphs. It is worth noting that information contained in text paragraphs is better conveyed by semantically recognized paragraphs. A variant of the solution has been proposed in which an additional edge is added to the graph composed of the text lines of a document, connecting specific lines of different text columns. This solution allows to solve the problem of recognizing semantic paragraphs of text. Moreover, the results obtained are close to those of physical paragraph detection.
Degree
thesis:*- Grantor dc:publisher
- Institutional Repository of Vilnius University
- Year dc:date
- 2023
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Bečalis, Gytis,
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:ELABAETD192845980&prefLang=en_US