I am designing a project based around the Kinect tracking someone’s hands and changing the audio based on their placement.
I have the Kinect mounted high up, pointed directly down. I am wanting it to read hand data as participants are reaching forward and touching something. The problem that I am running into is that the Kinect Chop only seems to read data if it sees an entire body. I am unable to get skeletal hand data since it is looking straight down, and can’t make out the whole body.
Is there anyway around this while keeping the Kinect in the same location?
Thanks so much.
Have you tried changing the Skeleton parameter to ‘Seated’?
EDIT: nm, I misread your post.
Is your Kinect even reading the visual input as a person? If so, you could try excising all the data points using a select OP and trying that way. Or, if you need, you could just test if there is movement detected in a certain vicinity (like within a certain UV threshold).
Thanks for your help.
I didn’t realize that the CHOP won’t start reading skeletal data until it sees a whole body (or a seated one). With just hands in the frame it won’t pick anything up. I’ll try what you have said. I might just have to rely on the TOP.
I haven’t worked with this much, but I have mostly heard about people using a strip of the TOP to do things like this. Obviously you can’t tell if it’s actually a hand in the strip, but I think this is the only way if the kinect can’t see enough body.