Motion Tracking: Waving Goodbye to the Past
Sci-Fi fans should remember the scenes from Minority Report, where Tom Cruise stood in front of a futuristic computer, moving files around with only his hands. The ease and speed with which he navigated through the interface for the files he sought seemed faster than your standard mouse clicks. This scene and others like it serve as the inspiration for the Armchair Interface. This new and exciting technology utilizes web cameras to translate gestures and hand movements into mouse control, providing a framework for what they call Hands-Free Human-Computer Interface. They may term this a hands-free device, but you won’t be eating nachos while surfing the web with this gadget anytime soon.
Jeffrey A. Delmerico and Daniel R. Schlegel seek not only to give birth to a new technology, they want to do it right. In their paper, they discuss various other types of tracking device research currently being conducted, including a processor-intensive version of their own tracking – one that may not be capable of being run in real time. Other gesture tracking technologies have been attempted with menu marking, in which a user would only be able to utilize menus when he places his hand in a specific region of the camera and does not permit wider access to the Graphical User Interface (GUI). The gestures do not yet translate easily to a computer so augmented reality is still a ways off.
Other attempts at providing a hands-free (perhaps in a truer sense) human-computer interface, have involved tracking the head, nose, or eye.
Head tracking involves tracking the user’s head in a plane in two- or three-dimensional space. There is often a device worn on the head, offering multiple points of light or highly contrasting color for the camera to track. Reference points are mapped on the front of the face/head, which can be triangulated based on proximity to the camera, allowing the system to figure out which general direction the user is facing.
Nose tracking is accomplished in various ways, some of which are similar to both head and eye tracking in their own, respective ways. Either the nose itself is tracked using sophisticated algorithms that are able to distinguish between your nose and your forehead. The tip of the nose is tracked in three-dimensional space, allowing the user to point with their nose to target their selection onscreen. Other common methods involve either head or eye tracking, where either the user’s head or eyes are tracked, and as they move through a combination of head movements, the user’s nose essentially points at the target on screen. While it may seem like both methods are versions head tracking, the two individual methods track completely different things.
Eye tracking uses sophisticated algorithms to determine where the eyes are looking. This kind of eye tracking is specifically called gaze tracking. Gaze tracking is fairly difficult to perform, and often requires heavy calibration or static head placement. Other less accurate forms of eye tracking involve systems of relative movement, where the speed and direction of the eye movements control a cursor. Gaze tracking is obviously the Holy Grail of eye tracking, as finding a way to remove the need for heavy calibration or static head placement would provide users with the ultimate hands-free method of targeting their selection on a computer screen. What could be more natural than merely looking at what you want to point to?
While these other methods of tracking may be more hands-free than, say, a gesture tracker that tracks hand movements, Delmarico and Schlegel are not allowing semantics to prevent them making their dream a reality. In the next part of this article, we explore the methods by which these two pioneers have expanded their fledgling technology and look at some of the lessons learned along the way.
Armchair Interface: A Framework for Hand Tracking for Hands-Free Human-Computer Interaction
Related articles: