Figure 
shows a few sample shots that the system
suggested during the course of normal play. These are the images
directly projected onto the user's eye. At this point, we are
undertaking a performance analysis of the overall system. The
reliability of the algorithm is being investigated as well as its
accuracy. In addition, time lag from input to output is currently on
the order of hundreds of milliseconds. Thus, the user has to maintain
smooth, slow head motion and occasionally wait for the system's
output. We are investigating techniques for making the algorithm more
efficient to avoid lag which is particularly distracting when the user
moves excessively. An evaluation of the effect of the system on the
user's billiards skill level is also to be performed. For instance,
how much better (or worse) does the user play depending on his or her
initial degree of competence? How much adaptation time is generally
required to see positive improvements? Another issue is offering
multiple styles of output display graphics to suit the particular
preferences of the user. Finally, we wish to evaluate the user's
non-augmented performance after he or she has been 'trained' with the
Stochasticks system. It would be interesting to see if the user
acquires a greater visualization ability as a result of the augmented
reality experience and to analyze the change in hand-eye coordination
due to the head-mounted display.
