![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
Meeting tomorrow
Feb. 3rd, 2005 10:52 pmTomorrow's my meeting with one of the people from the HIT lab at UW. (I know who, but I'm not saying specifically, because I'm nervous!) I've been thinking about all the material again for the last couple of days, and re-skimming the most important papers. (Mmmmm, vagueness!)
I think I'll be able to explain what I want to do, to research. I just wonder whether they'll buy it.
I did rediscover something I'd forgotten before - multiplexing images into segments, individually presented in extra-rapid sequence, seems to work. Which makes sense, but it's more strongly implied in the papers I have than I'd previously realised. (I suspect he knows it works but didn't want to say so - he was about to go found a startup, after all.)
It also occurs to me that if I want to try some sort of brainwave-controlled interface, a simple early way to manage the sensor array might be to build them into chobit ears. People would just think it was cosplay.
Or a fashion statement. Who knows? ^_^
I don't really think that's a long-term way to go, but it might be helpful in the early stages. Plus, it'd be cool.
I've been doing some math off the Intellectica paper - it might not be unreasonable to extrapolate a 220x220 output matrix, assuming no advances over their (vague) numbers. (Maybe 300x160 might be more appropriate...) That might just be a lot of data channel right there. That'd similarly be cool. Not to mention much more l33t than what I've gleaned from the patent data.
Anyway, I should post this and think more about Things (tm).
( And here, have quizzes! )
I think I'll be able to explain what I want to do, to research. I just wonder whether they'll buy it.
I did rediscover something I'd forgotten before - multiplexing images into segments, individually presented in extra-rapid sequence, seems to work. Which makes sense, but it's more strongly implied in the papers I have than I'd previously realised. (I suspect he knows it works but didn't want to say so - he was about to go found a startup, after all.)
It also occurs to me that if I want to try some sort of brainwave-controlled interface, a simple early way to manage the sensor array might be to build them into chobit ears. People would just think it was cosplay.
Or a fashion statement. Who knows? ^_^
I don't really think that's a long-term way to go, but it might be helpful in the early stages. Plus, it'd be cool.
I've been doing some math off the Intellectica paper - it might not be unreasonable to extrapolate a 220x220 output matrix, assuming no advances over their (vague) numbers. (Maybe 300x160 might be more appropriate...) That might just be a lot of data channel right there. That'd similarly be cool. Not to mention much more l33t than what I've gleaned from the patent data.
Anyway, I should post this and think more about Things (tm).
( And here, have quizzes! )
