I’m trying to use a kinect2 to detect people passing in front of a wall and use the silhouette to produce particles and other effects however I notice that there is a delay of a few seconds between the moment you pass in front of the sensor and the moment you get the silhouette feed.
Is there a way to reduce this delay/trick by mixing different methods to get a silhouette right away?
I tried to use a camera feed and use cache/blur/threshold to get it, but even if it is way faster, the result can be too messy in the end to be used in that case.
while a delay is expected, a few seconds sounds like something is wrong. Is the video feed from the Kinect delayed by a few seconds or is it that the effect only kicks in after a few seconds?
The video is not delayed, the fps of the effect in touchdesigner is around 55 fps, and in kinect studio the recognition seems to be faster.
Currently the player index shows up barely before the kinect bones (to give you an idea of the delay)
which is a problem as people can pass in front of the wall before the detection happens. However if one person is detected, any one entering the field of vision of the camera is almost detect instantly and if the person leave the field of vision and re renter within few seconds, there is no delay and the Kinect shows the player index right away…
It is really strange…