im very skeptical of the accuracy claimed. The layout and complexity of objects in most homes to do this is way to awkward to work reliably.
For someone breathing or a heartbeat you need much higher GHz signal. Usually this is done at 30ghz to 60ghz. The power flux leaving the antenna has an inverse square drop off rate which makes this basically impractical unless your standing directly in front of it.
I do agree some of these claims are pretty extreme. I wonder if it does work what the reliability of things like breathing and heartbeats. FWIW, some of these systems do incorporate 60GHz signals in their analysis, but as you mention dealing with 60GHz is incredibly challenging even in something like a residential building.
I'd really like to actually see it in person to really grasp the limitations.
Is 60GHz not part of the standard now? Only a matter of consumer hardware support.
I have personally tested wifi imaging from a cheap old 2.4Ghz linksys router that was accurate enough to tell if my hand was open or closed, maybe 10 years ago.