Hacker News new | past | comments | ask | show | jobs | submit login
Artists Create “Infinite Skulls” Using GANs (artnome.com)
56 points by Artnome on Feb 7, 2019 | hide | past | favorite | 8 comments



LSD and DMT always made me see "infinite skulls" so immediately I was drawn to this HN submission.

IMO, aticles explaining AI art processes are a great medium for introducing non-CompSci folks to AI concepts, like how this piece does regarding GANs - the gallery exhibit even more so.

One of my Comp Sci professors used a very similar technique for the designs he had printed on 4'x4' canvases which hung from the walls in his office.



My first time I looked at an artexed ceiling and it looked like hundreds of Jimi Hendrix face but in the style of Edvard Munch's scream. Surprisingly beautiful


> LSD and DMT always made me see "infinite skulls" so immediately I was drawn to this HN submission.

In case you missed it, this is a good article:

https://www.theatlantic.com/technology/archive/2015/09/robot...


Pretty cool! Excited for the day GANs produce something more meaningful though. I hope that after enough misfires like this eventually someone will produce art that is compelling on its own, without being carried by the "ai" label.


For what it's worth, we're using a GAN to generate fake user avatars for our products. They're real enough that we can use them in advertising, and since there's no actual person whose photo was taken, we don't require a signed model release form - something that was really difficult to get out of modeling studios since we wanted to open source the images after. GANs really solved our problem in this sense.


This would be a great addition to a service like https://placeholder.com/

They could offer GAN placeholders for many different subjects, like landscapes, cars, groups of people.

If you run such a service, let's talk about the implementation :)


It'd be really interesting to try that. I think anything organic it'd work well with - people, landscapes, etc. But I've noticed that a lot of the cars that come out of stylegan are pretty abstract - it seems to have trouble with machined objects. Lines don't always line up, which is fine on something organic since there's already an aspect of randomness in our bodies, but on mechanical objects it feels foreign.

Once we open source this I'd be interesting in doing gan placeholders as a service though, completely license free.




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

Search: