I feel as though this particular project is a never ending quest for perfection, which rather embodies the feeling of “The Fool” tarot card design which I have been working on.
As per my prior post, I opted to cancel my MidJourney subscription, but that will not stop me making the most of the thing before it expires properly. Then, it will be full tilt creation via Stable Diffusion.
In any case, I have probably created way too many images on this concept, and it is time to try and refine the outputs and short list what I believe is going to be the final design for this card and for this piece.
First, I’ll have to figure out how long ago it was since I started this endeavour in order to go back in time and grab all the candidate images. This is no easy task, I've generated over a thousand images using MidJourney at this stage, and it will take a very long time to download and extract the compressed files into a directory where I can identify the image by the prompt used.
I might have to resort to some photo-shoot style image culling, and that is quite the okay thing to do, in my opinion. :D (If time consuming)
Looking through all these images, there's one that really seems to be the most appropriate for the theme of the fool. It appears to be this one.
It needs a number of improvements, however. Firstly, upscaling to a higher resolution to allow print, and then there's the matter of the extra hand that appears to be emerging from the hip of the character.
Finally - The face could do with some more coherence, and the arms are a bit wonky, too.
This is likely going to be done via a combination of Stable Difusion and Photoshop, as I want to learn some new skills along the way. I'll get to that in another post, as reviewing all of these images took a considerably longer time than I would have liked.
The downside is that I have so many other images to review and enhance, and then card borders to worry about, too.
Why am I doing this? Because it is a fun challenge, and well, because I feel like I have to extremely prolific in order to feel useful.
Want more content from me?
Witness my futile efforts to play my Steam Game collection in alphabetical order.
Are you aware that I love photography? Check out my work in a collection.
If you want to see my Splinterlands antics and rants live, Find me on Twitch
If you prefer sleeping in your designated time zone, go watch replays on YouTube.
This looks cool... I want see all this photos 👀👀👀👌🔥🍷🍷🍷
Thank you
Your pictures are really great but I notice that on this picture many have the top of the head cut off. Are they whole at the beginning? or is it by placing it here that you had to shrink them? or is it deliberate?
The heads being cut off is an artefact of the way in which the AI generates these images. Since it is trained on 512x512 sample images, when generating an image that is greater in dimensions, it will often cut off body parts.
This can be fixed by generating 512x512 images, then doing what is called out-painting, to expand the canvas and its content.
Ok for the outpainting I'll have to try.
Other users have also suggested this: third person view, seen over the shoulder. maybe also a description of the hair would allow to have the whole head. i have several tracks to explore now.
Thank You
Interesting. I would never think about using third person view in the prompting but an interesting way to get around the issue! Thanks for sharing.
13yo popped in to find me staring with an odd expression and came to have a look, and we both agreed with where it cropped it looks like she has two elbows on the near arm.
An absolute monster in terms of proportions, yes. Can it be fixed? Maybe. Can I fix it?
insert terrified laughter here...
Sounds like sibling dearest the other day when she was trying to fix some photos that had been taken with the wrong lens XD
Try fixing it using Stable Diffusion! You only need a monster GPU!
She does not have a monster gpu XD