by johno » Wed Apr 23, 2008 8:32 pm
Indeed...that is the paper on which the tracking algorithm is based. The full algorithm is implemented in assembly in the file CamInterfaceAsm.S, which I tried to heavily comment so that it made sense (I usually hate reading other people's asm code).
If you get the algorithm running in an FPGA, I'd love to hear about it. This is something I've always thought about doing, but haven't had the time to pursue it. Things really get interesting when you can hook two cameras up to the FPGA, which would provide the ability to calculate disparity maps between the stereo cameras allowing for determination of depth.
Good times...if only there were more hours in the day.