alvr-visionos
alvr-visionos copied to clipboard
Face tracking (for OSC etc)
Ideas:
- Run an ML model on the Persona camera feed
- Submit Feedback until Apple adds a proper API
- Double-check Apple's sandboxing (might involve unsafe code that wouldn't pass app review lol)
It looks like the persona camera feed might be the best bet for the short term. Is it possible to obtain that feed and hide/process it while also rendering?
It looks like the persona camera feed might be the best bet for the short term. Is it possible to obtain that feed and hide/process it while also rendering?
not possible to hide it, which is why I've kinda lost motivation to look into it. I mean you can get the feed without rendering it, but it requires a window to be present, and the persona is rendered facing out from the window (if you turn 180° it will render the back of your head), and Apple doesn't allow repositioning windows for security reasons.