Forum Rules,TOS and Privacy Policy | By visiting this website and viewing the pages within, you accept and agree to be bound by and comply with our Rules Of Conduct, Terms Of Use and Privacy Policy
i am hoping to incorporate a web cam eye for my a.i. doll (would have to be wireless of course . ) that will work with some sort of software with my computer i have asked the ultra hal forum about any programs . what i want is a decent priced program that will recognize objects with info. put into the file ssystem and pics of colors and objects . if it would work the way i want if the video cam (in the doll) saw me in a blue shirt the program would recognize it by color and the shape of a shirt and the doll would say something like "thats a nice blue shirt you have on (and my name ( as it recognized me from my picture in the file too. so does anyone know of any object recognition programs that will work on xp ???
I dunno... That kind of software is still in the experimental stage. Most people who have it are probably working on it behind closed doors.
You may get lucky so do a search on google. You may need some programming ability in the long run tho.
I was at SIGGRAPH a few years ago and saw a demo of it. A video camera was targeted at spinning shapes in an aquarium. The computer could track them by both color and shape.
The type of AI you are looking for is called "computer-vision" or "machine-vision". You will likely find a good amount of information using those as a search on google. I don't think there is anything that will work exactly how you need without doing a lot of C++ coding. If you are good at programming, you will likely be able to write your own software using a basic image recognition library and extrapolating the details you need for specific cases (like a blue shirt or whatever). You might also look into facial-recognition, this will be essential for AI (telling when the user is smiling, for example).
It uses a neural network to process images. You could, for example, store reference images of different people (like a well dressed man or a sloppy bum) and have the AI determine if an arbitrary feed displays a gentleman or a bum. There is some example source code provided as well, it may something to start toying with. That same site also has a ton of tutorials on machine vision (in the articles section), which should be able to give you a good idea of what is out there. I hope that helps.
I would love to have that source code Cyber so I could hook it up with my karigirl project.
But you know... I don't see how its all supposed to come together without different people in the AI and robotics field coming together. I have my own field (which is natural language processing and awareness principles) and then there is something like Image recognition and things I have no time for. We are so engrossed in our fields because they represent new territory.
Its gonna take some doing for us to get to the next step if we want that robogirl. Grassroots...
check this link..the guy is no webdesigner but dig around a lilttle on his research into object recog. and robotics.. I think you might find what you need. http://sluggish.uni.cc/