Thank you all for your responses.
I had read those articles comparing the different notations, and I have also been aware that there is a lot more Sign Writing data available than anything else. Thank you for the direct links to things, I will take a closer look and do some more guided exploring. I plan to spend the next week doing some more in-depth research in this, and to play around with some code and some data. I will get back to you all when I've managed to give this the appropriate focus it requires.
Thank you Arthaey, Adrean. I'd be interested to include si5s in my resarch.
I would be glad to provide a review copy of How to Write American Sign
Language if you are interested, Jessica.
In your request you asked for a machine-readable file of the writing
systems you listed -- right now written ASL (si5s) does not have a
full font, although we are working on it. Once we get it right, I'm
sure that it will help immensely with machine translation.
Another thing to consider: Chinese/Japanese is able to be written on
the computer using a trackpad, as well as on the tablet. Perhaps
something like it could be altered to receive written ASL input. Are
you open to tinkering or only require ready files to be studied?
I'm also happy to provide translated sentences, depending on the time
limit and amount needed.
On Mon, Sep 3, 2012 at 6:10 PM, Arthaey Angosii <[log in to unmask]> wrote:
> Hi Jessica,
> If you're doing a comparison of written sign language systems, I would
> find it really interesting if you would also investigate "si5s". I got
> a review copy of the latest book from Adrean Clark (CC'd), but I
> haven't had time to compare it with SignWriting yet. Perhaps Adrean
> would be willing to give you a review copy?
> On Mon, Sep 3, 2012 at 11:37 AM, Jessica Hutchinson
> <[log in to unmask]> wrote:
>> Hi everyone
>> My name is Jessica Hutchinson. I am a Computer Science/Linguistics
>> postgraduate student at Rhodes University, South Africa, doing a research
>> project in machine translation of sign languages. My research involves
>> looking at the various notations used to represent sign languages and making
>> a comparison in terms of how effectively each of them can be parsed by a
>> computer. The results of my research I hope will further improve the area of
>> Sign-To-Text translation.
>> I would like to build a couple of small systems that do statistical
>> translation from sign notation to English text, for each of the main
>> notations (starting with Sign Writing, HamNoSys and Stokoe). To do this,
>> however, I need quite a bit of data, and I have had trouble finding any kind
>> of parallel corpus for these notations. I am hoping to do the translation on
>> a sentence-by-sentence basis, so the kind of data I am looking for would be
>> sentences in a sign language matched to sentences in English. I'm expecting
>> ASL to be most accessible, and for comparison's sake would prefer each
>> notation in the same language, but anything will be useful.
>> Furthermore, data would need to be in some kind of format that a computer
>> would be able to process - images in a pdf file aren't going to work so
>> If anyone can help me to find that data that I need, I would greatly
>> appreciate it.
>> Jessica Hutchinson
>> Research website: http://www.cs.ru.ac.za/research/g09h2318
Support American Sign Language through writing!