Re: deaf-uk-bsl again

— In, Alison Bryan <alison.bryan@g…> wrote:
> As a few people have asked me aboout this, I've temporarily changed
> the archive settings for deaf-uk-bsl …. to public.
> Archives here:

Hi Allison,

Thanks for the peek! This is such a wonderful idea. I
wish I understood BSL (or ASL, for that matter)! From
my point of view, the titles of your posts are a sort
of thread summary of a particular conversation. I can
see why some people are very interested in the
"metadata" problem for video. Ideally, you'd like to
be able to search on the content of video, too.
Because in your case, the content is language! (I'm a
computational linguist, by the way, so this is really
exciting stuff to me.)

There are definitely people (at universities) working
on this problem for content-based video retrieval…
but I doubt they imagined people using video the way
you are. This stuff is pretty far out… you'd most
likely combine gesture understanding with a standard
such as MPEG-7 which provides the means for encoding
metadata for content indexing and retrieval.

In the near term, I guess you already know, your best
bet is manual annotation of videos. Ideally, you want
a tool that can automatically segment your video into
smaller chunks which you can then easily markup with
useful descriptions or key words.

I'm really enjoying using a camera phone that allows
me to capture tiny pieces of video. I'll write a post
on this very shortly… but the nice thing is that
once your webpage loads, all of the video is
immediately accessible… there is no waiting time for
video to load since the pieces are very small —
around 150kb for 15-seconds. This seems really
applicable to the video wiki idea…