A camera fixed to the ceiling watches the activity
on the board. It is connected to a computer which
uses low-resolution images to detect when the user
has stepped aside and so is not obscuring the
board. The system then captures a high resolution
image, thresholds it, and attempts to recognise the
marks that are in the image.

My Ph.D. thesis subject was Video-Augmented Environments, in which I looked at ways of using video input for the creation of novel human-computer interfaces. Video cameras cost little more now than keyboards and mice, and there are plenty of interesting things you can do with them, but at present it often isn't worth the programming effort. I built an architecture which allowed you to create simple 'triggering' circuits from basic components (sources, sinks, filters & sensors) using a scripting language.