Novel AI for Vore Game Art

[wutterumeen]

Another tip would be to add the tag 'High Quality 3D render' tag to stuff on the furry option to avoid some of um, Microsoft paint looking pieces the AI tends to produce on that setting.

[bfdsjkbjjjj3443]

The way I understand AI art is that the AI takes art made by humans and combines multiple images together to make a new image. This is why there are copyright problems.

I would like to be able to see all the human made art that the AI used to make the new image, so that I can look at the human art and AI art next to each other and compare them.

[Content2020]

How do you guys get mawshots with inner detail? What I'm doing is
" (anything)girl, solo focus, looking at viewer, open mouth, wide open, inside throat, close-up, squinting,"
This gives me either a black chasm or just a flat colour. I get an unstable render if I try to add more specifics.

[Kavvanshrike]

I quite like the idea that ai art be used to help streamline some of the art assets made in some of the vore rpg maker games here on Eka's. Both MMA and GC tend to utilise photomanipluation of existing anime art to effectively add bellies and warp orifices for vore but are ultimately limited by the suitable images they can find and ai image generation kind of raises that limit by alot through taking requests. Personally I'd like to see how Salamoun or Mysta(also maybe ryanshow) uses ai image generation as a tool the way it SHOULD be used. By that I mean use an ai image as a base image in photoshop/gimp, remove the background edit hands and feet and fix any incoherencies with their photomanipulation abilities and edit in larger bellies,struggle, bumps, internals other characters being crammed into orifices as necessary for vore scenes.

[SaioKisaragi]

I quite like the idea that ai art be used to help streamline some of the art assets made in some of the vore rpg maker games here on Eka's. Both MMA and GC tend to utilise photomanipluation of existing anime art to effectively add bellies and warp orifices for vore but are ultimately limited by the suitable images they can find and ai image generation kind of raises that limit by alot through taking requests. Personally I'd like to see how Salamoun or Mysta(also maybe ryanshow) uses ai image generation as a tool the way it SHOULD be used. By that I mean use an ai image as a base image in photoshop/gimp, remove the background edit hands and feet and fix any incoherencies with their photomanipulation abilities and edit in larger bellies,struggle, bumps, internals other characters being crammed into orifices as necessary for vore scenes.

That is exactly why I made this post. I think this could be a great tool for people that want to make vore games. It removes some of the limitations the current vore assets have and lets these creators tell a story or build a game they really want! I hope this can streamline the process as well and make it easier and more accessible for more people to put their own creative spin on a vore themed game.

[Kerbalmaster]

Any Tips for making any Inside Stomach POV stuff? Like seeing inside her tummy?

[BellyHungerLady]

Would you take suggestions for new pics?

[Jaxed]

Would you take suggestions for new pics?

I might, but if you're wanting a specific character that isn't recognized by the ai, then there usually isn't much point in trying.

[Glorbgon]

Personally, I'd rather have bad/amateurish art in a game that something made by an AI. It makes me really uncomfortable whenever I realize I'm looking at AI drawings, kinda like when you realize you're in a dream because people around you look weird and off-putting. Also, since it's essentially a bunch of stolen art put into a blender, I much prefer to get games with straight up stolen art, Vore Days for example. Either way it's art taken without permission, so I'd rather not have it look like a shapeshifting alien trying to pass for a human being if I get to choose.

[Chessa]

The way I understand AI art is that the AI takes art made by humans and combines multiple images together to make a new image. This is why there are copyright problems.

I would like to be able to see all the human made art that the AI used to make the new image, so that I can look at the human art and AI art next to each other and compare them.

This would be a good experiment, but the AI doesn't work like that, and so the experiment is just a way to prove it doesn't.

I'd recommend looking into Stable Diffusion Single Image Finetuning, it is trained on that image and rebuilds it from the ground up to be entirely unique but with all of the qualities/themes intact, I think this is pretty close to what you're after and it proves that it actually understands the concepts well enough to apply them in unique ways and doesn't just copypaste multiple images.
https://www.youtube.com/watch?v=1TPdPBXNz7A <= This is a great example.

[Chessa]

Any Tips for making any Inside Stomach POV stuff? Like seeing inside her tummy?

Simple background, or grey background, crop out the character, place them inside belly, place belly ontop of characters belly, should be about the simplest way. Would require generating the pred, generating the prey, cropping the prey, possibly AI resizing the prey, using a belly image, layering the prey into the image, places on the pred.

[Eniel]

Do you think this algorithm can generate images for much more obscure characters that have very little or no pre-existing art at all?

[Jaxed]

Do you think this algorithm can generate images for much more obscure characters that have very little or no pre-existing art at all?

Generally no, unless they have a very simple design that the ai can be walked through using only descriptive tags.

[SaioKisaragi]

Novel AI Furry Diffusion has gotten even better!

Spoiler: show

[Scrypter]

That is good

[Gutlover]

Holy hell this is going to put vore artists out of work. Can it do vore acts though? Like swallowing or unbirthing or anal vore? What about different styles other than anime or furry? Has anyone tried realistic human preds with it?

If it can do custom imagesets I wonder if BigBig would be ok with putting their art through it. I'd love to see what it can do in that style. But only with their permission of course.

[Jaxed]

Holy hell this is going to put vore artists out of work. Can it do vore acts though? Like swallowing or unbirthing or anal vore? What about different styles other than anime or furry? Has anyone tried realistic human preds with it?

It can't do OCs, which make up the bulk of commissions, so I wouldn't be too worried.

[Chessa]

People worry, but artists will still be highly valued due to how precise their artpieces can be vs AI, and if they stylize in a way like Night in the Woods, Cult Simulator, or many other similar abstract styles, the AI can't really get it at all. Artists will likely even benefit in the long-run thanks to being able to generate a 'template' to start from and then modify the desired features on, etc. It could churn out lots of art at a normally backbreaking speed without nearly as much work.

[iforgot]

or if you dont want to pay money and you have time you could use stable-diffusion-webui,
https://github.com/AUTOMATIC1111/stable-diffusion-webui (dont be intimidated most of the time it should work with one or 2 copy and pasted commands)
https://rentry.org/voldy

Cacame on the UH discord got some really good results for Vore "If you have a decent enough GPU, you can find the leaked NovelAI models and decent tutorials on how to get everything running on 4chan, just go to vg or g and search for NovelAI or Stable Diffusion"

https://discord.com/channels/6367432365 ... 6528121003 (i think this should be the download link for a set of images that Cacame made)

[Rrevo]

or if you dont want to pay money and you have time you could use stable-diffusion-webui,
https://github.com/AUTOMATIC1111/stable-diffusion-webui (dont be intimidated most of the time it should work with one or 2 copy and pasted commands)
https://rentry.org/voldy

Cacame on the UH discord got some really good results for Vore "If you have a decent enough GPU, you can find the leaked NovelAI models and decent tutorials on how to get everything running on 4chan, just go to vg or g and search for NovelAI or Stable Diffusion"

https://discord.com/channels/6367432365 ... 6528121003 (i think this should be the download link for a set of images that Cacame made)

Link leads to an empty channel/server, atleast for me.