Mounting a 3D camera above the desk allows us to capture subtle motions of the hands and fingers.
Since its introduction in 2010, the Kinect has made 3D camera technology available to ordinary consumers, opening up a new world of gestural interfaces for playing games. However, the existing Kinect software and algorithms work best when capturing large, full-body actions for users that are several meters from the sensor.
Our technology enables 3D cameras based on the Kinect to reconstruct a finger-precise representation of what the hands are doing. This allows us to build simple and intuitive interactions that leverage small, comfortable gestures: pinching and small wrist movements instead of sweeping arm motions, for example.
Our system provides millimeter-level accuracy of the user's hands using a camera (PrimeSense Carmine 1.09) mounted about 70cm (28") above a desk.
Because the camera is mounted above the desk, it is still possible to use our system even when the hands are very close to the desk surface.
Because the camera is mounted above the desk, our system works well even if the hands are just a couple centimeters above the keyboard or desk surface, avoiding the so-called “gorilla arm” problem. In most cases, you will find you can rest your forearms comfortably on the desk surface while making small motions with your wrist and fingers. We even (experimentally) allow writing on the desk surface itself using your index finger (see our video).
Robust tracking at your desk
Our system is designed to fit on top of your desk and work well alongside the mouse and keyboard. It's also designed to run in the backround all day, recovering gracefully from tracking failures, such as when you step away from the desk.
Application Programming Interface (API)
Our API is open, meaning that you can write software in any language that takes advantage of it. It's free for academics, hobbyists and small commercial entities; contact us for a license if you are interested in integrating our API into your product. We have provided a number of example applications, and will be actively improving the tracking and the API in response to user feedback.
Get started building a gestural user interface.