Antw: [SLLING-L] Computers interpreting ASL

[Franz Dotter]

Dear Christian,

You may understand my harshness, when I tell you my experiences: Because technicians describe their wonderful results they want to reach by the strategy which can be also seen in this example (saying: we will solve all problems), I then meet politicians and administration people who tell me: We will not need interpreters or more teachers or more money for deaf education, because we soon will have FULL automatic recognition of sign language and even translation from signed to spoken and vice versa (and some technicians let the responsible persons stay in this opinion). By this play, some millions of research mone go to the technical side and e - from the social side - are not given the same mone to e.g. develop education and training materials for sign language users.

Therefore I demand that the technicians should not let themselves be "misunderstood" but they should simply say to the Technical Review: All what we want to reach in two years is to recognise some dozens of isolated signs and we will build a database to connect the recognitions and some items of another language, etc. If they will do this, we will not get into this asymmetric situation that technique is always seen as the science having the absolute, elegant and complete solutions while the social sciences cannot offer the same. Just: stay realistic in what you promise!

Best Regards

Franz


>>> "Christian Vogler" <christian.vogler at gmail.com> 01/13/09 9:48  >>>
On Tuesday 13 January 2009 10:28:50 Franz Dotter wrote:
> What the Boston colleagues have until now: An idea about how to recognise
> sign language (no, perhaps not sign language, only single gestures?) and
> that it would be useful to put that idea into a bilingual "dictionary" (no,
> not a dictionary, a word list). I'm in favour of technique, but, please
> stay realistic.

I think that you are being a little harsh here. The goal of this project is 
specifically not to recognize and transcribe entire sign language 
conversations and sentences, which is far more difficult that recognizing 
isolated signs (but progress is being made there, too, albeit slowly). The 
group is not claiming otherwise for their project, unless the article managed 
to distort the picture.

The goal is to be able to search a dictionary or video by sign, which is 
amazingly useful, not just for learners, but also in archive retrieval. Think 
of it - the thousands of sign language videos in libraries, as well as vlogs, 
are currently next to impossible to search, and unlike transcribing and 
translation, retrieval does not require you to interpret the sign first 
(although eventually it will help).

Best regards
- Christian

-- 
Christian Vogler, Ph.D.
Institute for Language and Speech Processing, Athens, Greece
_______________________________________________
SLLING-L mailing list
SLLING-L at majordomo.valenciacc.edu 
http://majordomo.valenciacc.edu/mailman/listinfo/slling-l


_______________________________________________
SLLING-L mailing list
SLLING-L at majordomo.valenciacc.edu
http://majordomo.valenciacc.edu/mailman/listinfo/slling-l