. Detecting, Tracking and Interpreting Faces: Real-time Tracking of Face Features

A   non-intrusive   real-time   program  is   developed  which  detects  the  eyes   and   nose   of   a  moving workstation user at a rate of between 10 and 30 Hertz. The program creates a base facility for other 

capabilities such as detecting gaze direction and facial  gestures, creating  face models, and normalizing for face recognition.  A skin color model is used along with geometric  knowledge about the face and weak  assumptions about the  lighting.  Good results are reported over various conditions,  including facial   hair,   3D   motion,   clothing   color,   and   use   of   eyeglasses.   Good   performance   has   been  demonstrated with dozens of subjects on a low end SGI workstation with an eye-camera to acquire images.

Our work is directed toward a general capability to detect and track a human face as it moves in a 3D workspace. Having  achieved this capability,  it can then be used to enable others.  For example, 3D pose can be used directly for Human-Computer Interface  (HCI) or for evaluation of how humans explore   computer   displays  or  virtual  environments.   An  extension  of   the  system  that  enables  to determine into which region on the workstation screen the user is looking is under development. The news system will enable to the user to issue commands to the computer using head movements and  gaze direction.