BracesForum.net
English => General => Topic started by: MarkMaster on 29. December 2022, 11:56:30 AM
-
Just a couple of questions that came to my mind:
* AI-generated images are the latest hype. Would you like to see AI-generated images (photorealistic or other) of people with braces? Or do you prefer real images?
I think this may help to solve one of our biggest issues - posting links to images without people's consent.
* For now, image generation algorithms seems to struggle to generate braces on teeth. Should we train a model with our vast collection of braces images?
I would love to see that, but I guess it would be a lot of work.
* ChatGPT can be used to write stories - what about an AI-generated braces story?
If I was a writer I would probably let an AI help me with my texts. As a reader I am curious what story ideas an AI would come up with.
What do you think? I'm looking forward to your comments.
Mark.
-
I like the idea of AI braces images as long as they are fairly realistic I'm not into anything to weird so for me yes I think they are another way to create images for a story especially where the braces are more of a fantasy rather than reality.
-
I wouldn't mind seeing pictures of fictional characters. Sometimes it would be nice if I could add a picture to one of my stories. I don't have the skills to create images.
As for computer-generated stories, I've seen a few examples of what ChatGPT can do. While it shows promise, the examples that I have seen tend to follow a similar script each time, with minor changes. As a writer, I don't know if I should be insulted that a computer program is being designed to try and replace what I can do. As a tool to aid me in writing stories, I'm not sure how I should feel. I already use spell-checkers and dictionaries, so having another tool to help me seems okay.
-
Good Points, M. :)
1. My attempts so far which were quite since Dall-e and Stable Diffusion came along remains the same even in the current version.
2. Without "fine tuning" the model (training with photos on this topic), it will take a while before it is possible to see a simple photo of someone smiling with retainer showing the palate, or wearing headgear attached to braces.
3. ChatGPT is very good in writing very appealing stories on braces fetish as long as you don't use trigger words related to explicit topics.
4. In fact, anyone can easily use it for sensual and interesting braces stories skeletons and there are some examples in the braces story section:
https://www.bracesforum.net/braces-stories-braces-special-(english)/open-ai's-gpt-3-got-upgraded/
Summary:
Does any of our subscribers have this technological abilities which are not trivial?
If not, maybe it can be bought as a service considering enough people want this, then it is also cheap.
Anyone?
-
Sorry but the idea really does nothing for me.
Personally I am only interested in either total reality or fiction which is very close to reality.
-
..... It takes art from real artists without their permission ....
Of course, you could argue that an artist could be similar.... I'm thinking about how an artist might "be influenced by"... that concept applies to ANY form of art... songs, paining, acting.
We, as non-artists do it all the time too, simply by "looking at stuff", we are then influenced, and it changes "what we like" and so on....
-
Crowded Princess,
I'm afraid what you wrote could be political or philosophical but currently, as you put it out, it is based on shallow, interest ($$$) based articles of many sources, with economical fears from power groups.
AI does not, needs not and never "copy". Specially visual art. its' ability to be influenced and directed towards specific art style or artist produces time after time novel creations that never repeat. If your text is descriptive enough, it is your text to image art.
And, with or without us agreeing, this tip of the iceberg will now never go away and the iceberg is surfacing faster one might think.
We rage when world order changes without warning, and specially when the change is not to are likings. Alas, I believe we will adapt and already are growing to surprising directions as humans.
-
I think, all copyright-discussion aside, we should absolutely try to train an embedding or hypernetwork for braces. I found no model so far that is able to generate anything but jumbled metal in peoples' mouths and this definitely has to change. Hit me up, if you think we could work on something.
-
Has anyone had any luck getting DALL-E to generate realistic images of headgear?
-
A young woman with braces standing in a living room. A thick silver wire extends from her braces, running over her face to her neck, passing between h
https://dereferer.me/?https://www.imagebam.com/view/MER3E4F
A young woman with braces. A thick silver wire extends from the braces, across her face to her ears. From there, the wire is connected with a rubber b
https://dereferer.me/?https://www.imagebam.com/view/MER3E4M
A young woman with braces. A thick silver wire extends from her braces, running over her face to her neck, passing between her lips. The wire is connehttps://dereferer.me/?https://www.imagebam.com/view/MER3E4O
-
No, this is not interesting for me at all.
-
It would be exciting to take the time to train your own image AI. You could feed the AI with many pictures of people wearing headgear... so that the AI thinks that's just how every person looks.
-
I can see the attraction of this have a lot of AI Generated images on my deviantart account none of my own work just copied over from other people's pages. Don't worry the original creators name us on each image. I have to confess I wouldn't even know how to start doing my own pictures but the idea is fun