Depth values from skeleton data differs from camera depth values

Hi,
I’m using nuitrack with RealSense D455. When tracking a joint of the person with nuitrack I get very high depth values (over 1000) but when I manually get the point in the view and the depth value corresponding to that point I only get values around 100.
Is there something I need to configure for it to be the same?
image

(Don’t mind the x and y values, I did not pick the exact same point, but it doesn’t matter, since the whole mapping seems to be off)

Hi @sassmannc,

but when I manually get the point in the view and the depth value corresponding to that point I only get values around 100.

Could you please describe how exactly you get the depth value of the point manually?

I did not pick the exact same point, but it doesn’t matter, since the whole mapping seems to be off

  1. Since it’s not exactly the same point, you may have selected a point that really has that value (100).

  2. The manually selected point could fall into the noise pixel/incorrect depth value (Nuitrack’s skeleton points are noise robust).

Hi @sassmannc,

How are you? Has your issue been solved?
Would be great if you could provide some reply/feedback, we will be ready to help.
Thanks.