Hello Everyone! First of all, I'd like to say I am extremely impressed and grateful for the roborealm software. This software can do anything!
I and a friend are building a robot we call Nina. I've built a few previous robots myself, and this is our first project that uses computer vision, so we're a little new to the science. I've been playing around a bit with RR and seeing if I can get it to work for Nina. Part of what I want Nina to do is be able to recognize faces. Iv'e tried this by using the skin pipeline and image matching for my face in particular, to see if it can just recognize me. But I can't seem to get it at a high enough match confidence. Also, it seems to recognize the walls around my house as being of a skin tone color.
I'd like to hear any suggestions on how to make roborealm recognize faces. What would be a plausible way to allow roborealm to recognize a broad range of similar objects, like whole personages?
I'm also trying to find the best way for roborealm to recognize soda pop cans and whatnot, but I'll leave that for the next issue, I guess...
Thank you!
|
|