Youtube:
Youku(The same as the u2 video above..):
It’s developed in C# using Kinect for Windows SDK 1.5.
I put the Kinect sensor high and tilted it down like this:
Now let me introduce my simple way to detect whether you’re gesturing like Gesture 1 or Gesture 2.
1st Step:
Limit the depth data you get. In my case, it’s between 950 to 1120. The red parts are what we get now.
2nd Step:
Limit the area you get so it only gets the area of your two hands.
3rd Step:
Only get the top side of the depth data we get. The red line is what we get now.
How to do this: I add the X coordinate of the depth data to a listbox if there’s no this such X coordinate in the listbox. And next time if you find that the listbox contains the X coordinate you don’t have to record it again, pass it.
4th Step:
If the count(length) of the listbox(the red line) is larger than a value(in my case it’s 30), you’re gesturing like Gesture One. Or you’re gesturing like Gesture Two. Because the Gesture One makes a longer red line.
That’s All~
Sorry for my poor English..