In Depth(Kinect OpenNI) is something written about transforming values. How does the following sentence look when it is patched:
In order for a pixel to be converted from projective-space to realworld-space you need to do something like:
realworld.x = (projective.x - 0.5) * projective.z * XYtoZ.x + projective.xprojective.z;
realworld.y = (0.5 - projective.y) * projective.z * XYtoZ.y + projective.yprojective.z;
realworld.z = projective.z;
where:
projective.xy = uv coordinates of a fullscreen quad
projective.z = value from the depthmap
XYtoZ = tan(FOV/2)*2
Are there any useful example patches?
Thanks
Armin