New Japanese Sign Language Dataset Aids Deaf Communication Tools

Researchers have developed JSL-DC, the largest Japanese Sign Language (JSL) dataset to date, featuring 36.7K videos from 19 signers. This dataset was created with a Deaf-centric approach, involving Deaf and Coda linguists in lexicon selection and data review to aid parent-child communication. The dataset includes linguist-derived descriptions to help differentiate between similar-sounding signs, and a model inspired by these descriptions achieved a 9.8% improvement on a confusable subset compared to existing methods. The JSL-DC dataset and its linguistic descriptions will be released under a CC-BY 4.0 license to promote research in sign language recognition. AI

IMPACT This dataset and its linguistic insights could accelerate the development of AI tools for sign language recognition, improving accessibility for the deaf community.

RANK_REASON The item describes a new dataset and associated research paper published on arXiv. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.CV →

• arXiv
• CODA
• Creative Commons Attribution 4.0 International
• deaf
• Japanese Sign Language
• JSL-DC

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →