Live data from Hacker News

AI real-time human full-body photo generator

generated.photos

101–110 of 421 posts

Re: AI real-time human full-body photo generator

#103
post #10

Earlier quoted context omitted.

There haven't been any solid rulings on the copyright validity of human-driven AI generation yet. There have been a few cases, but they've been muddied by complicating factors (not a human doing the generation - that is, autonomous generation - or the generation being used as a base work for something else). Additionally, even if there's no copyright, the terms of service may still apply separately (see OpenAI disall…

https://news.ycombinator.com/item?id=37195509 https://www.theverge.com/2023/8/19/23838458/ai-generated-art...

I believe that as the article says despite the headline this case was specifically about a situation where a computer scientist wanted to list the AI as the one creating the work. The case doesn't examine an argument that things can be copyrighted when a human is involved either by filtering the output or even just by developing the algorithm involved and thus the human is the artist and the AI is just a tool. I think what's clear is that legally AI can't itself create a copy righted work just like a camera can't be listed as the author of a work, but it's not clear if a human using AI as a tool either through prompting or filtering counts as a creative act under copyright or if AI generated creations count as derivative works of the models weights.

Re: AI real-time human full-body photo generator

#104
post #86
post #9

If you're wondering how it's so fast and cheap and they can generate variants so easily, it's because they're using GANs (see the footer). GANs are way faster than diffusion models because they generate the image in a single forward pass and their true latent space encoding makes editing a breeze. (And if you're wondering how it can look so good when 'everyone knows GANs don't work because they're too unstable', a wi…

>they're too unstable', a widespread myth >See for example BigGAN I remember when you try training a BigGAN model on anime images, the quality was bad. Now look at this example, one single GPU, 1.5M images with a diffusion model: https://medium.com/@enryu9000/anifusion-diffusion-models-for... ,the difference in quality is absurd, you can say this or that is not true but the quality speak for itself, obtaining good qu…

Broken link?

Re: AI real-time human full-body photo generator

#105
post #76
post #73

Earlier quoted context omitted.

Yeah, this was my first attempt: https://generated.photos/human-generator/64e648819c8c0400088... Not quite right. I am, however, impressed that the fingers are generally "mostly" correct.

Oof, I thought we banned thalidomide. I wonder if their pose detection/interpolation struggles for rarer poses, eg "kneeling with legs splayed leaning forward" is quite specific in saucy contexts and fairly sparse in more typical model shoots, so the manifold gets a bit holey, and overlaps with similar poses like one knee up, one hand forward.

Yeah I think that's it.

Also, it's way too easy to make something that looks like (or basically is) child porn with this tool. I chose Adult but something else keeps triggering it to generate a child-like face like in the image I posted. And as you showed, it's easy to get accidental NSFW pics.

Re: AI real-time human full-body photo generator

#106
post #37
post #27

NSFW

quite the opposite. I get a lot of the outputs filtered without any NSFW prompting. > We detected that generated image contains nude content. Try changing parameters.

oh, in a few iterations i got a nude sexy adult woman. clearly they're at risk of generating child porn (you can change the age to child or teen, though for obvious reasons I haven't tried it).

Re: AI real-time human full-body photo generator

#107
post #7

> “If you want to use images produced by Human Generator in commercial projects, contact us.” If there is no copyright in AI-generated images, then how can they possibly enforce this?

Its terms of use of the site, you are offered use of their geneator in exchange for agreeing not to use it commercially. If you use it commercially, you are breaking those terms, which they will argue are an enforceable contract. Copyright has nothing to do with it.

But what if you use the generator, post the image in an allowed non-commercial context, and then I copy that image and use it commercially? I have no contract with the AI generator company, and you didn't violate yours; it would seem to me that the violation involved is a copyright violation.

Re: AI real-time human full-body photo generator

#108
post #104
post #86

Earlier quoted context omitted.

>they're too unstable', a widespread myth >See for example BigGAN I remember when you try training a BigGAN model on anime images, the quality was bad. Now look at this example, one single GPU, 1.5M images with a diffusion model: https://medium.com/@enryu9000/anifusion-diffusion-models-for... ,the difference in quality is absurd, you can say this or that is not true but the quality speak for itself, obtaining good qu…

Broken link?

Fixed thx

Re: AI real-time human full-body photo generator

#109
Incredible. In the time I'd spend creating one image that exactly fits my or my client's needs or buying a high quality stock photo, I can generate literally millions of photorealistic, unappealing, images that would require a skilled commercial artist to make useful for all but the most throwaway uses. What about for some high-volume throwaway use case? I generated like 5 images before I got a 3/4 shot instead of a full-body shot. bzzzzzzzt.

Trying to 'wing it' with engineers doing what designers should be doing is a bad enough when you're just making regular interfaces, but when you're trying to sell a commercial art product, you need people with subject matter expertise. No matter how cool the technology is, and no matter how well it theoretically serves a commercial art customer base, if you're selling art, it's going to be critiqued as such. Hope you've got a thick skin.

Post reply on HN