i know that, every project is different, but by now i thought there must be a tried and proven path to do old fashioned (no kinect).
do you trust trautnerand contour, do you use external software like http://www.infomus.org/eyesweb_ita.php or some magic OpenCV plugin?
what i aim ot do:
camera (b/w everfocus, IR lit) sees 6x7 meter active area from above.
x-value of the rightmost person
a value for over-all activity (are people dancing or standing still)
and, at best:
xy-coords of each person to project a circle around him/her
(or: a hint to how the video -> shader approach works)
unfortunatelly the very promising node10 workshop on tracking by @frank ))and ((user:wirmachenbunt is “still” missing the patches :-)
thanks a lot for any help!!!