Elamur Research⌕
☰
home / papers / 2026 / Libras-UFPel Corpus: A Parallel Dataset of Brazilian Sign Language and Portuguese for Multimodal Research and Processing

Libras-UFPel Corpus: A Parallel Dataset of Brazilian Sign Language and Portuguese for Multimodal Research and Processing

Paper · 2026 · Antonielle Martins, Brenda S. Santana, Francielle Martins et al.

Summary

Sign Language Processing for Brazilian Sign Language (Libras) is limited by a lack of high-quality, interoperable parallel data. Existing sign datasets also tend to come from narrow scripted domains that underrepresent non-manual markers. The authors present the Libras-UFPel Corpus, which pairs Libras video with written Portuguese. It combines two sources: studio-recorded elicitations, in which native deaf signers built sentences around prompt signs, and naturalistic multi-camera interviews from the Inventário Nacional de Libras (INL-Pelotas). Annotation is done in ELAN using an ID-gloss scheme adapted from Johnston's guidelines. The scheme has a tier hierarchy that separates lexical, partially lexical (classifiers, pointing) and non-lexical (gestural) units, with independent tiers for facial and body cues. Pairs of linguists, at least one of them a deaf native signer, translate and validate the material. The current gold-standard subset of 1,200 sentences shows substantial lexical diversity, frequent classifiers and verb agreement. The resource is intended to support recognition, machine translation and future semi-automatic annotation, while linking descriptive linguistics with NLP.

Key points

  • The controlled subset holds 4,800 video recordings with Portuguese translations, split evenly between full sentences and standalone signs (2,400 of each). Four native deaf signers were recorded with a fixed frontal camera at 60 fps under uniform lighting.
  • The INL-Pelotas naturalistic component plans 18 paired interviews (36 participants, 3–4 hours each) filmed with four cameras. About 10 hours from three interviews are currently being edited.
  • The 1,200 sentences annotated so far contain 1,913 unique gloss types and 6,616 tokens, with a mean length of 6.3 signs (σ = 3.7).
  • Annotation conventions:
    • Lexical variants are indexed (e.g., CASA.1, CASA.2).
    • Prefixes mark classifiers (DV:), pointing (IX:), gestures (GES:) and fingerspelling (FS:).
    • Eyebrows, eye gaze, mouthings and torso each have their own synchronized tier.
  • The validated 1,200-sentence subset is planned for release through a tiered-access repository that protects signers' image rights. Human-in-the-loop annotation is planned to scale up the corpus.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

 Fields, topics and concepts

📚 Shelves 2The field of science this work belongs to (OpenAlex fields).Computer Science · 16,667 worksArtificial Intelligence · 16,194 works

🧭 Topics 1Research subjects and languages the work is about.Vision & language · 657 works

Cite

Cite as Antonielle Martins, Brenda S. Santana, Francielle Martins, Tatiana Lebedeff, Darley Nunes and Luisa Bohm. 2026. Libras-UFPel Corpus: A Parallel Dataset of Brazilian Sign Language and Portuguese for Multimodal Research and Processing. Proceedings of the 17th International Conference on Computational Processing of Portuguese (PROPOR 2026) - Vol. 1.

The paper itself, not this page. BibTeX from the library's record of the work.

@inproceedings{martins2026libras,
  title = {{Libras-UFPel Corpus: A Parallel Dataset of Brazilian Sign Language and Portuguese for Multimodal Research and Processing}},
  author = {Antonielle Martins, Brenda S. Santana, Francielle Martins, Tatiana Lebedeff, Darley Nunes and Luisa Bohm},
  booktitle = {Proceedings of the 17th International Conference on Computational Processing of Portuguese (PROPOR 2026) - Vol. 1},
  year = {2026},
  url = {https://aclanthology.org/2026.propor-1.112/}
}

References

19 references, 2 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Danielle Bragg and 1 others. 2019. Sign language recognition, generation, and translation: An interdisciplinary perspective. In Proceedings of the 21st International ACM SIGACCESS Conference on Computers and Accessibility, ASSETS ’19, pages 16–31. ACM. find
  2. Lucinda Ferreira Brito. 1984. Similarities & differences in two Brazilian sign languages. Sign Language Studies, (42):45–56. find
  3. Onno Crasborn and Han Sloetjes. 2008. Enhanced ELAN functionality for sign language corpora. In Proceedings of the 6th Language Resources and Evaluation Conference (LREC), pages 39–43, Marrakech, Morocco. find
  4. Mathieu De Coster, Dimitar Shterionov, Mieke Van Herreweghe, and Joni Dambre. 2024. Machine translation from signed to spoken languages: state of the art and challenges. Universal Access in the Information Society, 23(3):1305–1331. find
  5. José Mario De Martino and 1 others. 2023. Neural machine translation from text to sign language. Universal Access in the Information Society, pages 1–12. find
  6. Mirella De Sisto, Vincent Vandeghinste, Santiago Egea Gómez, Mathieu De Coster, Dimitar Shterionov, and Horacio Saggion. 2022. Challenges with sign language datasets for sign language recognition and translation. In Proceedings of the 13th Language Resources and Evaluation Conference (LREC), pages 2… find
  7. IBGE. 2010. Censo demográfico 2010. Instituto Brasileiro de Geografia e Estatística. Accessed: 2024-05-20. find
  8. Trevor Johnston. 2019. Auslan corpus annotation guidelines. Technical report, Macquarie University, Sydney. find
  9. Trevor Johnston. 2024. Auslan corpus annotation guidelines. Technical report, Macquarie University, Sydney. find
  10. Ronice Müller de Quadros, Jair Barbosa da Silva, Rodrigo Nogueira Machado, and Carlos Roberto Ludwig. 2020. Inventário nacional de libras. Fórum Linguístico, 17(4):5457–5474. find
  11. Ronice Müller de Quadros, Jair Barbosa da Silva, Miriam Royer, and Vinícius Rodrigues da Silva, editors. 2023a. Gramática da Libras, Vol. 1. Instituto Nacional de Educação de Surdos (INES), Rio de Janeiro. find
  12. Ronice Müller de Quadros, Jair Barbosa da Silva, Miriam Royer, and Vinícius Rodrigues da Silva, editors. 2023b. Gramática da Libras, Vol. 2. Instituto Nacional de Educação de Surdos (INES), Rio de Janeiro. find
  13. William C. Stokoe, Jr. 1960. Sign language structure: An outline of the visual communication systems of the american deaf. The Journal of Deaf Studies and Deaf Education, 10(1):3–37. Original work published 1960. find
  14. Peter Wittenburg, Hennie Brugman, Albert Russel, Alex Klassmann, and Han Sloetjes. 2006. ELAN: a professional framework for multimodality research. In Proceedings of the 5th International Conference on Language Resources and Evaluation (LREC), Genoa, Italy. ELRA.
  15. Kayo Yin, Amit Moryossef, Julie Hochgesang, Yoav Goldberg, and Malihe Alikhani. 2021. Including signed languages in natural language processing. In Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics (ACL), pages 7347–7360. Association for Computational Linguistic…
  16. Antonielle Cantarelli Martins, José Mario De Martino, Janice Gonçalves Temoteo Marques, and Francielle Cantarelli Martins. 2023. Construindo critérios de lematização para a língua de sinais brasileira. TradTerm, 45:147–179. Número Especial – Libras, Lexicografia e Cultura. find
  17. Robert Munro Monarch. 2021. Human-in-the-Loop Machine Learning: Active learning and annotation for human-centered AI. Manning Publications. find
  18. Medet Mukushev, Arman Sabyrov, Madina Sultanova, Vadim Kimmelman, and Anara Sandygulova. 2022. Towards semi-automatic sign language annotation tool: SLAN-tool. In Proceedings of the LREC2022 10th Workshop on the Representation and Processing of Sign Languages, pages 159–164, Marseille, France. Europ… find
  19. Zhe Niu, Ronglai Zuo, Brian Mak, and Fangyun Wei. 2024. A Hong Kong sign language corpus collected from sign-interpreted TV news. arXiv preprint arXiv:2405.00980. find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/martins-2026-libras-ufpel-corpus-parallel-dataset/