I'd be interested into introducing my current screen/virtual desktop as a source in the pipeline.
https://doc.qt.io/qt-5.11/qpixmap.html
With a configurable delay in ms (and virtual desktop number), it would allow, eg subtraction, of the (expected) screen output from camera-originated image of the projection of that screen.