~Red Dress~ by Ken1171_Designs
Close
Hover over top left image to zoom.
Click anywhere to exit.
Members remain the original copyright holder in all their materials here at Renderosity. Use of any of their material inconsistent with the terms and conditions set forth is prohibited and is considered an infringement of the copyrights of the respective holders unless specially stated otherwise.
Description
An experiment with Poser and Stable Diffusion, where I made a 3D render to use as an input for the AI to create a new interpretation. It has preserved much of the pose, proportions, and composition, just didn't get the shadow right. I asked it to add belts to the dress, and it did quite well. The end result is a mix of 3D render and illustration.
Comments (3)
EVargas
Interesting experiment! I see the knees got a strange curve on the outside, strange as it gets most of the main elements right and this line is just a long silhouette.
Ken1171_Designs
Thanks for the feedback. I need to spend more time on this to understand how to get the best out of it. The challenge is controlling the small things without affecting the entire image. When I got to an acceptable state, I used inpainting to rebuild the hands, which is almost always mandatory with SD. :)
MollyFootman
Interesting, for sure. Thanks for sharing! I still haven't spent much time with Stable Diffusion but I finally got it installed on my computer. Not all that easy for someone who has limited experience with the "guts" of applications only with their use. XD
Now I want to try giving it some input pictures after seeing this. As for hands..... >.<
Molly
Ken1171_Designs
Stable Diffusion keeps growing every day, and by now I think I am barely using a fraction of what it can do. Just last night I was told about a new extension called ControlNet that could had made this process much easier. I wish I knew that before I did this experiment. I have played with it for a few hours last night, and it introduces a timesink I will be dealing with. :)
Arodia
Pretty interesting! Seems like the shadow in the AI version is not really a shadow of her, but something else.
Ken1171_Designs
Yes, indeed the AI doesn't know that is a shadow, so it got it all wrong. I tried explaining in the prompt, but the AI still got it wrong. It has trouble handling more than 1 subject in the same image, I suppose.