Wish there was some AI tool that could do some proper ligne Claire. It seems to be quite impossible. Oh, as is Escher, but I do understand that one.
How a Stable Diffusion prompt changes its output for the style of 1500 artists
131–140 of 205 posts
Re: How a Stable Diffusion prompt changes its output for the style of 1500 artists
#132Very cool! I recently made a game kind of like AI "pictionary" where the user has to guess the "artist", subject, and description of a piece of art generated by stable diffusion: https://wcedmisten.fyi/project/paintingGuesser/ I tried to make something more general, but stable diffusion is fairly inconsistent in how well the output matches the semantics of the input.
This could be a good candidate for a wordle-style scoring system
No post body was provided.
Re: How a Stable Diffusion prompt changes its output for the style of 1500 artists
#133It is sort of funny/interesting -- I only tried a few, but famous anime or manga artists (try "Junji Ito" or "Hayao Miyazaki") seem to have at least one picture that is clearly the result of the algorithm picking up on their fans' art.
Re: How a Stable Diffusion prompt changes its output for the style of 1500 artists
#134This website should include a google-image-search link to the artist, so you can compare the styles more easily.
Re: How a Stable Diffusion prompt changes its output for the style of 1500 artists
#135It would be so much better to just have all of these 1500 pictures next to each other in a list with the artist's name under it.
Including a few real sample pictures from the artist, or a link to a google image search of the artist name, so you can check if the generated style matches.
Re: How a Stable Diffusion prompt changes its output for the style of 1500 artists
#136The concept of "derivative work" is pretty important in copyright law. I wonder if anyone has thoughts on this, in terms of this type of project. Should there be legal implications to this? I know someone -- a completely unknown artist -- who used to make a fair portion of their living by drawing D&D characters for people. Unfortunately, orders slowed down, because someone can input one of his images into software li…
"Should there be legal implications to this? ... Should this be allowed?" Even if there are laws against it, the cat's out of the bag. There's no stopping billions of people all over the world making derivative works at the push of a button.
> the cat's out of the bag.
Yeah, just wait until Disney characters get copied and mixed.
I wouldn't be surprised if their lawyers are preparing to change copyright law ... again ...
Re: How a Stable Diffusion prompt changes its output for the style of 1500 artists
#137Honestly I don't think its very good at emulating style. It picks up some things but often times misses the heart of the matter.
Did some spot checking with some of my favorite artists. Rockwell's paintings are all about storytelling, clearly not present in the work. Their emulation for frazetta doesn't look like frazetta's work at all. HR Giger emulation is a joke. David finch at least gets a penciler's style but misses the use of solid blacks and dynamic posing. Frank Miller doesn't look like miller's work at all. etc etc etc
This list goes on and on. Personally while I understand using the 'in the style of' as a way to change the image results, I think in many many cases the results just don't look like the art of that artist.
Re: How a Stable Diffusion prompt changes its output for the style of 1500 artists
#138Hate to say it but when i see stuff like this it only reminds me of what we could have achieved if this ingenuity had been applied in another domain. Can't help feeling that this accidentally harms creative types and risks swamping us with visual junk. The technical achievment is astounding but no-one would seriously claim that crafting an image via a short prompt is creative except in the most cursory way. I'm proba…
> only reminds me of what we could have achieved if this ingenuity had been applied in another domain. I hate arguments like this. Even ignoring how dismisive it is of the achievement at hand, why would you assume ingenuity is transferable like that? Someone who makes a breakthrough in physics is by no means likely to have made an equivalently ground breaking advance in biology if they had decided to study that field…
Aside from the fact that I explicitly praised the achievment, my point actually relies on said appreciation.
I guess my musing was hypothetical but I was careless in communicating that. I get that we can't centrally plan innovation or human effort - and I certainly wouldn't want to live in a society where this was the case.
Re: How a Stable Diffusion prompt changes its output for the style of 1500 artists
#139Interesting, how with billions of nodes and supposed "intelligence", the network hasn't been able to deduce a simple concept of symmetry in human faces. All of the eyes and all of the lips in all of the pictures are asymmetrical, which easily gives AI generated images away.
Many of the artists in this list specifically try not* to have this symmetry present in the faces.
It's what makes many styles of art separate from just taking a photo or simply going outside.
The system used here is actually astoundingly good at producing many artists stylesbecause it's not going for symmetry.
Re: How a Stable Diffusion prompt changes its output for the style of 1500 artists
#140I’ve been idly working on a similar list but with a much more basic prompt.
The results vary wildly, even run by run, but I might put them in a few buckets:
1. Similar enough that someone who doesn’t know much about art could be fooled
2. Amateur knockoff but recognizable style
3. Influence is there if you know what to look for
4. Artist probably not in the training data at all
The last one kinda surprised me, for artists whose work is online and who have unusual names. I would have thought those cases would be really good. Maybe they ran out of disk space with all the porn?
Also interesting that it gets much closer for figurative painters than for abstract painters.