Handling sign language transcription system with the computer-friendly numerical multilabels
This work addresses the challenge of processing sign language corpora for researchers and developers, though it is incremental as it builds on existing annotation systems.
The paper tackles the problem of automating sign language transcription by converting HamNoSys annotations into numerical multilabels, simplifying annotation structure without significant loss of gloss meaning to accelerate vision-based sign language recognition.
This paper presents our recent developments in the automatic processing of sign language corpora using the Hamburg Sign Language Annotation System (HamNoSys). We designed an automated tool to convert HamNoSys annotations into numerical labels for defined initial features of body and hand positions. Our proposed numerical multilabels greatly simplify annotations' structure without significant loss of gloss meaning. These numerical multilabels can potentially be used to feed the machine learning models, which would accelerate the development of vision-based sign language recognition. In addition, this tool can assist experts in the annotation process and help identify semantic errors. The code and sample annotations are publicly available at \url{https://github.com/hearai/parse-hamnosys}.