[kylebrown]

Wow.. I would love to develop for this.

I dropped out of state U. after my 3rd year (math major), but that was years ago. At my current start-up, I have recently been forced to learn much more than I was expecting to about probabilistic graphical models and curve similarity measures (gladly though; always been interested in pattern recognition).

Anyone with a vision for this, consider dropping me a line. I might be able to help.

[chime]

I'd love to integrate it with my KType application ( http://ktype.net ) or build a new software that converts hand gestures / ASL to speech. I think it's totally doable and would be a lot of fun technically too. What do you think?

[kylebrown]

Well, off the top of my head I don't see a straightforward way to integrate with KType (nice work btw!). Just because there would be so many possibilities for mapping hand gestures to the KType actions/commands.

On the other hand, converting sign language to text/speech seems like it should be quite straightforward. Not knowing anything about sign language, I'm assuming signs map (more or less) one-to-one with words. The input from LEAP appears to be extremely high resolution, so if the sign gestures are properly normalized (and judging from the demo video, it looks like the LEAP SDK itself already does a good degree of input normalization), you should be able to just train your classifier (neural network, SVM, etc.) right out of the box.

Of course, things are never as easy as they look so in all likelihood there are plenty of complications I'm completely overlooking at first glance. But I agree with you 100% that it sounds totally doable.

[btr41n]

So sign language isn't really one-to-one, but that is something that I really want to work with, so I look forward to getting a unit. I think this system has a lot of potential as a learning tool! Awesome job guys!

[ChrisFornof]

Chime, email us (community (at) leapmotion (dot) com) . We'd love to do what we can to encourage your research.

-Chris