Abstract
Sign language translation systems are complex and require many components. As a result, it is very hard to compare methods across publications. We present an open-source implementation of a text-to-gloss-to-pose-to-video pipeline approach, demonstrating conversion from German to Swiss German Sign Language, French to French Sign Language of Switzerland, and Italian to Italian Sign Language of Switzerland. We propose three different components for the text-to-gloss translation: a lemmatizer, a rule-based word reordering and dropping component, and a neural machine translation system. Gloss-to-pose conversion occurs using data from a lexicon for three different signed languages, with skeletal poses extracted from videos. To generate a sentence, the text-to-gloss system is first run, and the pose representations of the resulting signs are stitched together.
Original language | English |
---|---|
Title of host publication | AT4SSL 2023 - Proceedings of the 2nd International Workshop on Automatic Translation for Signed and Spoken Languages |
Editors | Dimitar Shterionov, Mirella De Sisto, Mathias Muller, Davy Van Landuyt, Rehana Omardeen, Shaun O'boyle, Annelies Braffort, Floris Roelofsen, Frederic Blain, Bram Vanroy, Eleftherios Avramidis |
Publisher | European Association for Machine Translation |
Pages | 25-36 |
Number of pages | 12 |
ISBN (Electronic) | 9789464857184 |
State | Published - 2023 |
Event | 2nd International Workshop on Automatic Translation for Signed and Spoken Languages, AT4SSL 2023 - Tampere, Finland Duration: 15 Jun 2023 → … |
Publication series
Name | AT4SSL 2023 - Proceedings of the 2nd International Workshop on Automatic Translation for Signed and Spoken Languages |
---|
Conference
Conference | 2nd International Workshop on Automatic Translation for Signed and Spoken Languages, AT4SSL 2023 |
---|---|
Country/Territory | Finland |
City | Tampere |
Period | 15/06/23 → … |
Bibliographical note
Publisher Copyright:© 2023 2023 VFS Handling Qualities Technical Meeting. All rights reserved.
Funding
This work was funded by the EU Horizon 2020 project EASIER (grant agreement no. 101016982), the Swiss Innovation Agency (Inno-suisse) flagship IICT (PFFS-21-47), and the EU Horizon 2020 project iEXTRACT (grant agreement no. 802774).
Funders | Funder number |
---|---|
Inno-suisse | 802774, PFFS-21-47 |
Horizon 2020 | 101016982 |
Innosuisse - Schweizerische Agentur für Innovationsförderung |