Process face detection on own samples / manipulating samples

[Sven_G_]

Hi all,

I'm new in Real Sense SDK and will try out something.
Working on streams / devices and recorded files works fine.
Is it possible to process the RealSense features (facedetection, objecttracking, gesture etc.) on own samples / imagedata?
I have tried the following, but it doesn't work:

while(true)

{
    sensemanager->AcquireFrame()

    sensemanager->QuerySample();
    sample->color->AcquireAccess(...)
    //fill image buffer
    sample->color->ReleaseAccess(...);

    ProcessData(...)

    sensemanager->ReleaseFrame();
}

I can manipulate the imagedata with that, but it has no effect on the essential process (f.e. facetracking)?
(I got the same problems with ImportData instead of Acquire-/ReleaseAccess)

Is it possible in generally and what do I'm wrong?