reverse facenet [closed]

asked 2018-07-31 01:29:48 -0500

holger gravatar image

updated 2018-07-31 01:30:47 -0500

Hello,

I hope this question is not too off topic. I have used a facenet model to get embeddings of a person, is it possible to (re)generate an image out of it(reverse processing)?

I want to do "draw a similar picture" for fun.

edit retag flag offensive reopen merge delete

Closed for the following reason the question is answered, right answer was accepted by holger
close date 2018-07-31 06:45:36.415293

Comments

1

@holger, take a look on GANs.

dkurt gravatar imagedkurt ( 2018-07-31 03:27:01 -0500 )edit

Hey - thank you again for the term, i was hoping for a something i can google on like last time :-)

holger gravatar imageholger ( 2018-07-31 05:52:46 -0500 )edit
1

maybe start with an autoencoder or Restricted Boltzmann Machine

berak gravatar imageberak ( 2018-08-02 02:29:58 -0500 )edit

Thank you for the terms too - i will look into it. These are all funny words somehow - like it ^^

Have a nice day + Greetings, Holger

holger gravatar imageholger ( 2018-08-02 04:23:50 -0500 )edit