Videos

AI Creates 3D Models From Faces | Two Minute Papers #149



Two Minute Papers

The paper “Photorealistic Facial Texture Inference Using Deep Neural Networks” is available here:
http://www.hao-li.com/Hao_Li/Hao_Li_-_publications.html
http://arxiv.org/pdf/1612.00523v1.pdf

Two Minute Papers Merch:
US: http://twominutepapers.com/
EU/Worldwide: https://shop.spreadshirt.net/TwoMinutePapers/

Earlier episode on texture synthesis: https://www.youtube.com/watch?v=8u3Hkbev2Gg
PatchMatch: https://www.youtube.com/watch?v=n3aoc36V8LM

WE WOULD LIKE TO THANK OUR GENEROUS PATREON SUPPORTERS WHO MAKE TWO MINUTE PAPERS POSSIBLE:
Andrew Melnychuk, Christian Lawson, Daniel John Benton, Dave Rushton-Smith, Esa Turkulainen, Sunil Kim, VR Wizard.
https://www.patreon.com/TwoMinutePapers

Music: Antarctica by Audionautix is licensed under a Creative Commons Attribution license (https://creativecommons.org/licenses/by/4.0/)
Artist: http://audionautix.com/

Thumbnail background image credit: https://pixabay.com/photo-1961529/
Splash screen/thumbnail design: Felícia Fehér – http://felicia.hu

Károly Zsolnai-Fehér’s links:
Facebook → https://www.facebook.com/TwoMinutePapers/
Twitter → https://twitter.com/karoly_zsolnai
Web → https://cg.tuwien.ac.at/~zsolnai/

Source

Similar Posts

36 thoughts on “AI Creates 3D Models From Faces | Two Minute Papers #149
  1. That scene in Enemy of the State where they rotate the interior of the lingerie shop to check if there was something in Will Smith's bag suddenly seems not like bullshit and actually realistic now. The computer doesn't even need two angles.

  2. I hope the people working on the paper get a job at the upcoming Mixed Reality companies (Microsoft, Magic Leap, Google, Facebook, Apple), Mixed reality (MR) is seen as the next best thing in computing with the biggest tech companies investing billions of dollars into it and yet not a single consumer MR device is out except for the big Oculus and Vive VR sets which don't count because they are not full MR.

  3. I was just looking at a user tutorial for zbrush and thought " what if you applied deep learning to this process" (the one in the tutorial) so I searched to see who was working on that (I'm not so egotistical to think I was first to the party on this, or much else for that matter) and here I am, this video is already two months old and the technology is being implemented in the digital modeling industry… It may have just sparked a revolution in Media production, manufacturing and indeed the long awaited VR gaming worlds of the future…which is now.
    Side note* super happy to see powerlaces from Nike now also a thing…we'll have hoverboards in a photorealistic Virtual world before we have them in this one at this rate. (Which is not such a bad thing if you ask me, personally).

  4. So this is just a guess but from what I understand, the neural network in question must be massive in size (I'm thinking in the hundreds of teras) and still learning to this day ! It gathers huge quantities of scanned data for both geometry and surface texturing (photogrammetry tech)… and based on the one picture you feed it, it smartly tries to guess the details from its own library. This is amazing, potentially it could applied to about anything knowing that skin is one of the hardest material to fool humans with, and as it was previously stated it does goes right through the uncanny valley !!! I guess it indeed explains why this tech is not available to the public ….

Comments are closed.

WP2Social Auto Publish Powered By : XYZScripts.com