This forum is disabled, please visit https://forum.opencv.org
 | 1 | initial version |
Do you know this paper from Andy Wilson? http://research.microsoft.com/en-us/um/people/awilson/publications/wilsonuist2006/wilsonuist2006.html
I implemented this pinch gesture with a kinct some years ago and (like wilson) used the center of the inner contour instead of the fingertips. If I remember correctly, I took the depth image, found the hole (your inner contour) and used a simple dilation (around 10px) of this contour as mask. By doing this, I got the 3d points on the border of the gap and used the mass of center as final Input Point. Andu Wilson proposes to do a pca on the pixels in the inner contour to get an orientation but this direction was rather instable in my experiments.
