Using zed-unity-livelink-fusion fused bodies are not registered unless pretty much all cameras detects the person very accurately. Even though I set minimum_allowed_cameras to 1 it still doesn’t register fused bodies. I can clearly see from the detection of the individual cameras that bodies are detected on multiple of them, however.
BodyTrackingFusionRuntimeParameters:
// define fusion behavior
sl::BodyTrackingFusionRuntimeParameters body_tracking_runtime_parameters;
// be sure that the detection skeleton is complete enough
body_tracking_runtime_parameters.skeleton_minimum_allowed_keypoints = 7;
// we can also want to retrieve skeleton seen by multiple camera, in this case at least half of them
body_tracking_runtime_parameters.skeleton_minimum_allowed_camera = 1;
In the OpenGL window of the sender (which you can enable here), do the skeletons tracked by only one camera appear? If so, do they appear thicker (fusion), or only thin (individual detection)?
I did a quick test on my side and could not reproduce your issue (with 3 cameras, only one camera seeing the person).
There was a bug in the main branch preventing fully functional multi-people tracking.
We released a fix just now, can you test it?
It’s possible that it’s not related though. How many cameras were you using?
If possible, can you send us a set of SVOs reproducing the issue, along with the Fusion configuration file?