Art by Artur-AE

[williej]

Hi Artur - As always, your work is very impressive and I'm interested in your recent work with AI. Apologies if you've mentioned it already in the thread but which app/program are you using to generate the initial images? Are you creating with Daz and then running your output through an AI image enhance program?

[wulf]

Enhanced Oldie from 2022 'In the Torture Prison of ESMA'

Another erotic masterpiece with one of my favorite devices... the hot seat.
I'm getting a very high reading on my eom once again, so thank you for sharing.

[artur-ae]

Hi Artur - As always, your work is very impressive and I'm interested in your recent work with AI. Apologies if you've mentioned it already in the thread but which app/program are you using to generate the initial images? Are you creating with Daz and then running your output through an AI image enhance program?

Process of creating an image (CG enhanced with AI)

Basically, I still create computer-generated images with DazStudio. These are then enhanced in the free Krea.AI*.
Later, I touch them up in PhotoScape X Pro. I use PhotoScape because it makes it easy to work on two levels.

I put the image to be edited (Krea.AI enhanced) in the “Editor”. Under the “Cut Out” tab, I load the original image created with DazStudio and cut out the objects to be exchanged (e.g. hands and tools), which I then insert into the image in the editor.

After that, the editor image is saved and in a second step in PhotoScape, the exchanged objects are then adjusted (with a mask technique). Once you have developed a routine, it is very quick and easy.




*Working with Krea.AI (you have to register for free):

I open the Krea.AI website. Link: https://www.krea.ai/home

Then I click on “Enhance” at the top of the command line.
A mask appears with a square in the center that says “Drag & drop a video or...”.
I drag the image to be edited into it.

Then a mask appears on the right with the image to be edited. The upscaling 2x is highlighted.
Do not change anything in this mask. Do not enter anything. Just press the “Enhance” button.

After a short time, the revised image appears in the middle. Right-click on the image to download it.

After the download, I delete the image on the left side again. Additionally, I open the images in the upper command line at “Assets” and delete the image there as well. You don't want to wake sleeping dogs.

That's all the work in Krea.AI.




Process of creating an image only with AI. (Before there are any more questions. )

Today, for images create exclusively with AI, I use the stability matrix from Stable Diffusion and the package “Stable Diffusion WebUI Forge”. With various flux checkpoints locally on my PC. The advantage here is that you have direct access to CivitAI or Hugging Face. For downloading checkpoints and Loras.

ComfyUI I use as a package on the same matrix to work with different workflows or to create new ones. For example, to cut out people in pictures or create small movie scenes. But I'm still practicing that.

This all happens locally on my computer, so it's uncensored.

[hornythumb]

This all happens locally on my computer, so it's uncensored.

Makes for optimal images.

[frog]

Basically, I still create computer-generated images with DazStudio. These are then enhanced in the free Krea.AI*.
Later, I touch them up in PhotoScape X Pro. I use PhotoScape because it makes it easy to work on two levels.

Thanks for the tutorial. I will try AI again with yours and JD's imput. Failed miserably the first time. All that talk of loras and work flows just befuddles me. Must be my senile brain.

In any case, I still don't really like what most AI "art" looks like. It seems the producers just throw out what ever the thing produces and do no further work. It all seems so uniform and alike. This is, of course, not the case with what you do. Just expressing what I've noticed.

[williej]

Many thanks for your comprehensive guide Artur. Very helpful indeed.
I will go and try your technique.

[artur-ae]

Hi Artur - As always, your work is very impressive and I'm interested in your recent work with AI. Apologies if you've mentioned it already in the thread but which app/program are you using to generate the initial images? Are you creating with Daz and then running your output through an AI image enhance program?

Process of creating an image (CG enhanced with AI)

Basically, I still create computer-generated images with DazStudio. These are then enhanced in the free Krea.AI*.
Later, I touch them up in PhotoScape X Pro. I use PhotoScape because it makes it easy to work on two levels.

I put the image to be edited (Krea.AI enhanced) in the “Editor”. Under the “Cut Out” tab, I load the original image created with DazStudio and cut out the objects to be exchanged (e.g. hands and tools), which I then insert into the image in the editor.

After that, the editor image is saved and in a second step in PhotoScape, the exchanged objects are then adjusted (with a mask technique). Once you have developed a routine, it is very quick and easy.




*Working with Krea.AI (you have to register for free):

I open the Krea.AI website. Link: https://www.krea.ai/home

Then I click on “Enhance” at the top of the command line.
A mask appears with a square in the center that says “Drag & drop a video or...”.
I drag the image to be edited into it.

Then a mask appears on the right with the image to be edited. The upscaling 2x is highlighted.
Do not change anything in this mask. Do not enter anything. Just press the “Enhance” button.

After a short time, the revised image appears in the middle. Right-click on the image to download it.

After the download, I delete the image on the left side again. Additionally, I open the images in the upper command line at “Assets” and delete the image there as well. You don't want to wake sleeping dogs.

That's all the work in Krea.AI.




Process of creating an image only with AI. (Before there are any more questions. )

Today, for images create exclusively with AI, I use the stability matrix from Stable Diffusion and the package “Stable Diffusion WebUI Forge”. With various flux checkpoints locally on my PC. The advantage here is that you have direct access to CivitAI or Hugging Face. For downloading checkpoints and Loras.

ComfyUI I use as a package on the same matrix to work with different workflows or to create new ones. For example, to cut out people in pictures or create small movie scenes. But I'm still practicing that.

This all happens locally on my computer, so it's uncensored.

Follow-up.

Stable Diffusion WebUI Forge is not stable at the moment. Prompts are no longer executed accurately. Completely different images than requested appear.
The following warning appears in the Stability Matrix: py:2: DeprecationWarning: pkg_resources is deprecated as an API.
This is somehow related to Python. You should port to importlib-based solutions.
This is too complicated for me at the moment.

I switched to Forge UI. It is also equipped with the same Stable Diffusion input masks. But it is based on A1111. Python is included directly in it.
But it works very well with Flux checkpoints and Flux Loras.
Unfortunately, you can't access CivitAI and Hugging Face directly.

[artur-ae]

“Death by Torture”

This is the first image created with ForgeUI and the checkpoint: fluxUncensoredFemale_v10.
Except for the prompt and the negative prompt, no Lora's have been used. I think it's a remarkable implementation of the prompt. The writing on the edge of the photo should actually be “Death by Torture”, but since no Lora was used for the lettering, the result is still good.

For those who might be interested, here are the prompts etc.

Prompt: a polaroid Photo with a txt written on the egde: "death by torture", of (a completely naked woman with long curly red hair, pale skin, very tiny breast, very fear and shocked expression), is standing with arms over head, hands on the neck, in a dark torture room like the 80s in argentina, beside her a (wardener with black beard with angry face in a black uniform),

Negative prompt: bad anatomy, bad hands, missing arms, extra hands, extra fingers, bad fingers, extra legs, missing legs, poorly drawn face, fused face, worst feet, extra feet, fused feet, fused thigh, extra thigh, worst thigh, missing fingers, long fingers, extra eyes, huge eyes, amputation, cartoon, cg, 3d, unreal, animate, cropped, blurry

Steps: 20, Sampler: Euler, Schedule type: Simple, CFG scale: 1.01, Distilled CFG Scale: 3.5, Seed: 3825705058, Size: 896x1152, Model hash: 17dfe52589, Model: fluxUncensoredFemale_v10, Version: f2.0.1v1.10.1-previous-636-gb835f24a, Module 1: ae, Module 2: clipLFineTuneByZer0int_text, Module 3: fluxT5XxlTextencoder_v10

[williej]

Just thought I'd post an early try using your tutorial Artur. I'm quite pleased with it.

[artur-ae]

Just thought I'd post an early try using your tutorial Artur. I'm quite pleased with it.

Red%2030-enhanced lo res.jpg

Five Stars. *****

Well Williej,

except for the slight disfigurement of the hands, it looks very good. And as is sometimes the case with Krea.AI, the faces have no Asian touch. It really looks good. Congratulations. A really great picture.

But you should still look into Fooocus.AI. Fooocus makes it easy to add new faces with new facial expressions. This is sometimes necessary if the ladies look too Asian, or have lustful looks or just laugh despite the boob whipping. With a little practice, that's also totally straightforward. Your first Enhancing is really something to be proud of.