In our previous entry, we presented the ability of the Kinect to create 3d map reconstructions but now, it can also recognize objects and provide feedback. A user can teach a PC to recognize objects and store them in memory using the Kinect. As seen here in this video by yankeyan, he uses the Kinect to detect the desired object. By using voice commands, the user is then able to register the item accompanied with the label given by the user. If the material is successfully labeled, the machine can recognize the item and notify the user from henceforth.
The developer used the following for the construction of this project:
OpenCV (image processing and recognition)
FestVox (speech synthesis)
CMU Sphinx (speech recognition)
One has to wonder though, what if you pair this technology with the 3d map reconstruction program and then use the autonomous navigation technology of Kinect, what would you get?