Quote: "and then tell me that is can't accurately differentiate between body parts.
"
All the demos are done standing up. What about people sat down or slouched? (potentially obscured by background objects etc.)
The silhouettes are clear when stood up, hence the perfect accuracy - but is that the same if the silhouettes aren't completely clear? Can the camera and depth camera work together to understand self-intersecting limbs, for example?
I wasn't "blindly" criticizing it: I was simply not being "blindly" swayed by what is effectively a marketing demo
EDIT: Presumably it uses motion to detect the person vs. the background - so what happens when multiple people are moving within camera view but only one person is playing? Does it continually perform facial recognition to work out who is playing?
It's really interesting technology, it would just be a bit more useful if they released some better detail on how it works (and how it will work practically in real households)
09-f9-11-02-9d-74-e3-5b-d8-41-56-c5-63-56-88-c0