Prompting with Sign Parameters for Low-resource Sign Language Instruction Generation

Fuente: arXiv
Enregistré dans:
Détails bibliographiques
Auteurs principaux: Tariquzzaman, Md, Ishmam, Md Farhan, Muna, Saiyma Sittul, Hasan, Md Kamrul, Mahmud, Hasan
Format: Preprint
Publié: 2025
Sujets:
Accès en ligne:
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866916917339488256
author Tariquzzaman, Md
Ishmam, Md Farhan
Muna, Saiyma Sittul
Hasan, Md Kamrul
Mahmud, Hasan
author_facet Tariquzzaman, Md
Ishmam, Md Farhan
Muna, Saiyma Sittul
Hasan, Md Kamrul
Mahmud, Hasan
contents Sign Language (SL) enables two-way communication for the deaf and hard-of-hearing community, yet many sign languages remain under-resourced in the AI space. Sign Language Instruction Generation (SLIG) produces step-by-step textual instructions that enable non-SL users to imitate and learn SL gestures, promoting two-way interaction. We introduce BdSLIG, the first Bengali SLIG dataset, used to evaluate Vision Language Models (VLMs) (i) on under-resourced SLIG tasks, and (ii) on long-tail visual concepts, as Bengali SL is unlikely to appear in the VLM pre-training data. To enhance zero-shot performance, we introduce Sign Parameter-Infused (SPI) prompting, which integrates standard SL parameters, like hand shape, motion, and orientation, directly into the textual prompts. Subsuming standard sign parameters into the prompt makes the instructions more structured and reproducible than free-form natural text from vanilla prompting. We envision that our work would promote inclusivity and advancement in SL learning systems for the under-resourced communities.
format Preprint
id arxiv_https___arxiv_org_abs_2508_16076
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Prompting with Sign Parameters for Low-resource Sign Language Instruction Generation
Tariquzzaman, Md
Ishmam, Md Farhan
Muna, Saiyma Sittul
Hasan, Md Kamrul
Mahmud, Hasan
Human-Computer Interaction
Computer Vision and Pattern Recognition
Sign Language (SL) enables two-way communication for the deaf and hard-of-hearing community, yet many sign languages remain under-resourced in the AI space. Sign Language Instruction Generation (SLIG) produces step-by-step textual instructions that enable non-SL users to imitate and learn SL gestures, promoting two-way interaction. We introduce BdSLIG, the first Bengali SLIG dataset, used to evaluate Vision Language Models (VLMs) (i) on under-resourced SLIG tasks, and (ii) on long-tail visual concepts, as Bengali SL is unlikely to appear in the VLM pre-training data. To enhance zero-shot performance, we introduce Sign Parameter-Infused (SPI) prompting, which integrates standard SL parameters, like hand shape, motion, and orientation, directly into the textual prompts. Subsuming standard sign parameters into the prompt makes the instructions more structured and reproducible than free-form natural text from vanilla prompting. We envision that our work would promote inclusivity and advancement in SL learning systems for the under-resourced communities.
title Prompting with Sign Parameters for Low-resource Sign Language Instruction Generation
topic Human-Computer Interaction
Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2508.16076