Reproduce “On the Importance of Sign Labeling: The Hamburg Sign Language Notation System Case Study” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Maria Ferlin, Sylwia Majchrowska, Marta Plantykow, Alicja Kwaśniwska, Agnieszka Mikołajczyk-Bareła, Milena Olech, Jakub Nalepa
Investigating inconsistencies in HamNoSys annotations across five multilingual sign language corpora, this study from the HearAI project found that subjective interpretation and lack of clear rules hinder their utility for training robust machine learning models for sign language recognition. The analysis quantifies variability within HamNoSys location classes and assesses the impact of symbol reduction.
Key insights
- Current HamNoSys annotation practices are plagued by inconsistencies and subjective interpretations, especially regarding the initial sign frame, dominant hand distinction, and hand location.
- Despite its linguistic utility, HamNoSys, in its current annotated form, is not optimally suited for direct computer-aided synthesis and recognition without standardized procedures or modifications.
- A significantly reduced set of HamNoSys symbols can still efficiently map back to the original gloss (word), suggesting a simplified representation might be sufficient for some automated translation tasks.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.