Our main goal for the head tracking, is to render the virtual world
in perspective with the positions of the eyes. If the eyes are moving
the perspectives of the virtual world must be changed in a way the
person thinks he is looking into a "virtual window". To make a
virtual world we use a view frustum. A view frustum is the region of
space in the modeled world that may appear on the screen. The frustum
is a pyramid from the view point to the far plane. The pyramid is
truncated at the near plane, hence the name frustum. Only the things
between in the pyramid between the near and far plane is visible from
the viewpoint. In the figure below you see a scheme of the frustum.
\begin {center}
  \includegraphics[width=99.7mm]{img/frustumscheme.png} \\
  Figure 5: A scheme of a frustum.
\end {center}
In the next figure you see the objects which can be viewed by the
view point. The green object can be seen totally, the yellow
partially and the red object can't be seen.
\begin {center}
  \includegraphics[width=81.5mm]{img/frustumobjects.png} \\
  Figure 6: The object which can be seen by the viewpoint.
\end {center}
 If the viewpoint is moving, the frustum
will be different. We use the the glFrustum function of the OpenGL
library to set the frustum at his new values. This is a function who
multiply the current matrix with a
perspective matrix. \\\\
The specification of glFrustum is as follows:
\begin{verbatim}
void glFrustum( GLdouble left,
                GLdouble right,
                GLdouble bottom,
                GLdouble top,
                GLdouble zNear,
                GLdouble zFar )
\end{verbatim}

The parameters left and right will specify the coordinates for the
left and right vertical of the near plane. The parameters bottom and
top will specify the coordinates for the bottom and top horizontal of
the near plane. The zNear, zFar Specify the distances to the near and
far depth clipping planes.

To calculate the left, right, bottom en top we use the coordinates of
the persons eyes. How the coordinates are retrieved and calculated to
the world values can be read in the previous chapter. We now set a
frustum with the near clipping plane at $(-1/2 world width, -1/2
world height, 0), (1/2 world width, 1/2 world height, 0)$ with the
eye at position \emph{(Eye Coordinate X, Eye Coordinate Y, Eye
Coordinate Z)}. But because OpenGL expects the eye at position
\emph{(0,0,0)} when specifying a frustum, we have to specify our near
clipping plane with \emph{(- Eye Coordinate X, - Eye Coordinate Y, -
Eye Coordinate Z)} as its origin, and translate the modelview matrix
to \emph{(- Eye Coordinate X, - Eye Coordinate Y, - Eye Coordinate
Z)} , such that the world origin is in the center of the screen. In
order to allow objects to appear in front of the screen we move the
near clipping plane closer to the eye position, without changing the
frustum. Below you see the demo application of the head tracking
technique.
\begin {center}
  \includegraphics[width=100mm]{img/HeadTrackScreenShot.png} \\
  Figure 7: The head track demo application.
\end {center}