![]()
Visiting guest. Why not sign in?
|
|
Going with the Flow Optical flow estimation for gesture recognition | |
Going with the flow
I've written a small demo program to calculate optical flow from an attached webcam. The program works quite nicely and zips along at 20fps on my 1.7GHz P4. At present it's written using the only C++ compiler I have to hand, Borland C++ Builder version 4, and I've got no idea what DLLs you would need to run the executable standalone since the help system is pretty vague on that score. Anyhow, the general idea is to use optical flow information to help my Rodney robot recognise certain types of gesture. The prime candidates are head nodding/shaking and finger/arm pointing. These elementary social events together with other information from speech recognition all help to give the robot important clues about what is going on and where it's attention should be focused. If anyone wants to have a go at porting the program to MFC or KDE I'd be interested to have a look at the results, since I havn't really done much C++ programming for years. - Bob |
|
Opticality
Congratulations on your 100th post Bob, I guess that makes you a special person ;) Pretty much anything would zip along on a 1.7Ghz machine... i really need to upgrade! You should be able to pickup a very cheap version of VC6 now, since MS' latest compiler is out... It's a very helpful page describing the technique you've put up. Nice work. Have you noticed any significant behavioural improvements in the little Rodney thanks to this? |
|
Visual perception
I guess 100 posts makes me an old timer. At the moment the optical flow estimation isn't integrated with the rest of the robot - it's just a standalone demo. But potentially this sort of information could be very useful not only for categorising different types of movement but also possibly for estimating the shapes of objects. I did read a paper recently where they had used stereo optical flow information to reconstruct partial 3D models. http://www.ri.cmu.edu/projects/project_464.html I think it should be fairly easy to reproduce the Cog head nodding experiment using a combination of skin tone recognition and optical flow. It might also be possible to have the robot exhibit a startle response if an object is moved quickly towards the cameras, which produces a "star field" type of flow pattern. - Bob |
|