Spaces:
Running
a little research
made a little observation to find out which render is better. When using hyper render and render, the spheres do not have an ideal shape, some of them are deformed, but for that hyper render gives them variety and the quantity is respected, as it was said in the request. On the other hand, when using super render and ultra render, the almost ideal shape of the sphere is observed, and only ultra render displayed the exact request regarding the amount, and also decorated them a little. If there are any other useful tips on the render hint, I will be glad if you share.
This is very interesting, I believe it has to do something with the AI using stronger words like "ultra", "hyper", etc.
Do you have idea why are almost all spheres independently on the render type, placed over the surface? I would expect just a ball and simple one colored background...
I tried to create porous structures...
I really like the two porous beads on the back on the second image. But I can not find the right key words to generate them exclusively :-(
I tried to create porous structures...
I really like the two porous beads on the back on the second image. But I can not find the right key words to generate them exclusively :-(
try hint texture
Cool, thank you very much! Sponge, holes, craters ... indeed, I have to go "out of the box".
Meantime I partially succeed with this
:
I do not know... The simplest answer will be to blame AI bias. AI was trained on data from the Internet and internet is full of Instagram-Queens :-) With adding texture you avoid details where AI fails as e.g. face.
I do not know... The simplest answer will be to blame AI bias. AI was trained on data from the Internet, and internet is full of Instagram-Queens :-) And with adding texture you avoid details where AI fails as e.g. face.
He didn't understand faces before.:)) if you add a hint, digital art, digital painting, artwork or fan art, as well as a beautiful face, everything will become much better :)
You are right, but any small imperfection on the face makes whole image ugly. Even minimal asymmetry of eyes, hands, body, etc. disturb the final outcome. For us, human beings, symmetric body/face = nice body face. For me, with adding texture, AI do not "understands the anatomy" of the body better, but it is avoiding (covering) details what in final makes the body/face more symmetric, i.e. nicer.
You are right, but any small imperfection on the face makes whole image ugly. Even minimal asymmetry of eyes, hands, body, etc. disturb the final outcome. For us, human beings, symmetric body/face = nice body face. For me, with adding texture, AI do not "understands the anatomy" of the body better, but it is avoiding (covering) details what in final makes the body/face more symmetric, i.e. nicer.
here I agree, but let's call it a kind of crutch so that it would not be so scary to look at people, I hope that soon the developers will resolve their issues and update the AI so that it can already show good results without any texture
Thank you very much for the help. I am trying to get the image that I could use for research project presentation, poster or graphical abstract. I made white porous beads, which are coated with fluorescent green coat and used for blood purification. Something similar to dialysis. I am tired to see PowerPoint spheres etc. and of course I am not so skilled to draw it by myself :-)
Thank you very much for the help. I am trying to get the image that I could use for research project presentation, poster or graphical abstract. I made white porous beads, which are coated with fluorescent green coat and used for blood purification. Something similar to dialysis. I am tired to see PowerPoint spheres etc. and of course I am not so skilled to draw it by myself :-)
I am very glad that I could help you, I wish you success in your presentation :)
Cat1- these are super cool!