SignWriting List Forum| Message 4666 | Previous | Next [ Up Thread ] | Message Index |
|
SignWriting List Forum |
|||
|
|
From:
Antonio Carlos da Rocha Costa Date: Thu Mar 29, 2001 6:40 pm Subject: sign language processing and computational sign language linguistics | ||||||||||||||||||||||||||||||||
|
Dear Valerie and Themis, > Also a question for the list members: Does anyone uses SW or .sgn files > for computational linguistics of sign languages? > > Thank you > Themis. That's the main purpose of the research project SignNet (Adapting Internet technologies for Sign Language Processing) that I'm running since 1999, here in Pelotas (UCPel) and Porto Alegre (PUCRS and ULBRA). Classically, there are two main research areas dealing with the subject of computers and languages: Computational Linguistics and Natural Language Processing. Briefly, the first is centered on using computer techniques to ease the study of natural languages. It's mainly a field for linguists. The latter is centered on developing computer programs to process natural languages getting hints from linguistic theories on how to do that. This is a field mainly for computer people, specially those working in the area of Artificial Intelligence (which is my case). Classically, also, both areas work with written language. That is, the objects they deal with are written words, sentences, dialogs and texts. And they have developed a large amount of techniques to deal with such things. In the case of sign languages, the usual approach people are taking is to work with either with the capture of signs in real time (as they are performed) or with videos, trying to extract sign features from real images. From the computational linguistics and natural language processing point of view, such works are not able to make use of the whole set of techniques that those areas have already developed, because they are directed toward language cast in written form. What SignWriting brings on behalf of "sign language computational linguistics" and "sign language processing" is exactly the chance of making those techniques immediately available to the work with sign languages. The SignNet project has as its main goal to develop software and techniques to be used as a basis for programs dealing with signs in written form (in SignWriting). It's particularly interested in developing techniques for sign language processing, that is, sign language interpretation. But not only this: software for dictionaries of written signs and databases of sign language texts where signs are structurally represented (and not only pictorically represented), for sign language animation, for sign language spellchecking, etc. Sure computational sign language linguistics seems would be near to us, and we hope that the sofware and techniques we are deloping be useful also for such area (although we have no experience in computational linguistics to tell if there is any real chance for that hope). All the best, Antônio Carlos | ||||||||||||||||||||||||||||||||
|
|