Sonoplastic
Technically I started with one of the most straightforward ways to do video sensing: performing frame-differencing with a mean filter which gives a reading of the amount of movement detected. Recently since the Leap Motion device is available I integrated it in my system (thanks to aka.leapmotion) to have a wider and better control. Using these techniques and tools within the Max/MSP/Jitter environment i am able to map the amount of motion in the scene where my hands are moving to specific algorithms that i conceived and keep enhancing.