Hacker News new | past | comments | ask | show | jobs | submit login

There's another thing from Snowcrash I've been thinking about - realistic faces. The most exclusive club in the Metaverse was The Black Sun and what made it special was that the avatars had extremely realistic, life-like expressions, enabling them to visually express emotions and thus enabling a higher form of communication (correct me if I'm wrong, I haven't read it in a decade).

So if we buy into the notion that social presence in VR worlds will be big, somebody's going to have to build this. The technology exists, but it looks very awkward: https://www.fxguide.com/wp-content/uploads/2012/12/HBT-VFXpr...

Not only does that have to be translated into a consumer product, but you need to capture someone's face while they're wearing the VR headset, which makes it even harder. Since VR headsets already touch your face, I would imagine the product would have to be some kind of extension of that - a larger contact area filled with sensors that reconstruct your expression perfectly.




High Fidelity [1] (Second Life v2) does this already. It's still early beta, but they've got the ability to map a user's face in real time with enough data-density to transmit emotion, and facial characteristics good enough to recognize a person by their avatar.

You are correct, that currently the same camera that does the facial analysis (macbook camera) precludes the VR headset - so you either get to see another in full VR, or you get to transmit your expression in HD. The device to do both is not yet here.

[1] http://highfidelity.io/


The founder of Second Life gave a live demonstration of this last night at the Silicon Valley Virtual Reality meetup. I didn't realize how important capturing these small facial gestures was until I saw it myself.


Link?


Facebook has thousands of pictures of our faces under a variety of lighting conditions. It would be exciting (or scary, depending on your point of view) to see that data, combined with sensors on the headset itself, to generate realistic-looking face models in the virtual space.


I just came here to post this. They also recently hired Yann LeCun to head up their AI Lab and also published a ground-breaking face recognition paper. So, they have the data, and they have the people.


I remember Hiro's ex-girlfriend Juanita explaining that when she designed the facial-expression system, she used him and herself as the models. So they are they only two people who have completely honest, natural expressions in VR.


I recently came across this effort to map human faces to avatar faces: https://www.facerig.com


I still can't believe that that's real-time from just a webcam.

This won't work as-is without cameras inside the VR goggles, of course, but details, details...


The Kinect already attempts to reconstruct your expression with the IR/image overlay. It's not the most accurate, but that could just be the software.


Go watch 'capturing avatar'




Join us for AI Startup School this June 16-17 in San Francisco!

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: