General AI Portraits

Welcome to our Community
Wanting to join the rest of our members? Feel free to Sign Up today.
Sign up
Masato Toys

Masato Toys

...
Jan 21, 2015
3,255
6,053
LOLOL




I don't understand how this works?? Do they have a massive database of portraits and just match facial recognition? Maybe imitate the composition by cropping?

I can't believe its generating portraits but these are uncanny lol
 
Masato Toys

Masato Toys

...
Jan 21, 2015
3,255
6,053
If anyone gives a shit, I binge watched like 4 seasons of this show last week. Most full season episodes are available on YouTube

Its really amazing but the judging sucks lol

Lots of fascinating artists I would like to have a beer with.

All the sitters are famous and pretentious pricks

good show
 
Belobog

Belobog

First 100
First 100
Jan 14, 2015
759
1,445
Masato Toys said:
seriously, how the hell is this being done?

Any smart people wanna take a stab?
Click to expand...
They trained Generative Adversarial Network (GAN) models to reproduce human portraits, with different styles and levels of abstraction. GANs are a very popular class of deep generative models. They are trained to learn a mapping of a latent vector z ∈ Z to a generated image y = G(z) with G being the generator. The latent space Z describes all possible portraits. AI Portraits Ars pushes them towards an intuitive and playful way of interacting with state-of-the-art GAN models. By showing a face to the neural network, they walk through the Zspace and identify the vector that best describes our face in the multidimensional space of the GAN. They trained AI Portraits Ars using their GAN on 45,000 portrait images. To allow insertion of own images into the latent space of a model, they trained an inverter that can approximate the latent vector z = I(x) from an image x.
 
Masato Toys

Masato Toys

...
Jan 21, 2015
3,255
6,053
Belobog said:
They trained Generative Adversarial Network (GAN) models to reproduce human portraits, with different styles and levels of abstraction. GANs are a very popular class of deep generative models. They are trained to learn a mapping of a latent vector z ∈ Z to a generated image y = G(z) with G being the generator. The latent space Z describes all possible portraits. AI Portraits Ars pushes them towards an intuitive and playful way of interacting with state-of-the-art GAN models. By showing a face to the neural network, they walk through the Zspace and identify the vector that best describes our face in the multidimensional space of the GAN. They trained AI Portraits Ars using their GAN on 45,000 portrait images. To allow insertion of own images into the latent space of a model, they trained an inverter that can approximate the latent vector z = I(x) from an image x.
Click to expand...
Yeah that's what I thought too (well, something similar), just wasn't sure
 
Last edited:
Grateful Dude

Grateful Dude

TMMAC Addict
May 30, 2016
8,929
14,278
Haha, this is some wild shit.

I wanted to try this out, so I used a photo of a MMA legend.

I present to you a portrait of the great Don Frye!

 
You must log in or register to reply here.
Top Bottom