Image generation around an existing PNG

[sgalissie]

Hi all, Do you know the best way to use DALL·E 3 for image integration? How can I adjust my prompt to ensure that DALL·E 3 fully incorporates the details of my uploaded PNG image into a bigger new one?

[mitchell_d00]

Welcome to the forum it is an intriguing place
There are a lot of variables in this question Dalle TOS blocks some types of images. But if you are just asking about prompt work to describe it. Use a standard logic, keep your prompts exact, tell the gpt to send your prompts exactly as you type them and to say exactly what was sent in image summary. Consistency in format I found to be incredibly useful in image prompts.

This thread by @Daller has became invaluable to me.
As has this one started by @PaulBellow. Both are almost a sub forum here.

Collection of Dall-E 3 prompting bugs, issues and tips

Yes this is probably the reason for the “Mouthy” template effect. but instead to over-train the system for faces, what they should do is give faces a extra level of details, specially around the eyes and the mouth, or all the face if they are small. Something like a area of up-scaling. MJ has fixed this problems, if i see images from it. Very very realistic faces, even if they are small.

DALLE3 Gallery for 2023/2024: Share Your Creations

Quack! Rubber ducks stacked to make a world of rubber ducks in a duck shape [IMG_3682]

[mitchell_d00]

And if you want to join in on some fun this thread is really wonderful. It is the spooky scary thread

Official DALL-E Gallery: October 2024 (Halloween Edition)

Hey Community, After much behind-the-scenes discussion, we’ve decided to give the DALL-E galleries a fresh spin. The showcase thread has become overwhelming, with many excellent creations buried beneath a TON of scrolling. Forum Regulars and Leaders have been going back and forth, discussing what we want from these new threads. We’ve considered everything from one-shot generation competitions and difficult challenges to problem-solving threads where the community helps each other with generati…

[_j]

The above information does not directly address the heart of the issue: misunderstanding of the technology.

@sgalissie

Do you know the best way to use DALL·E 3 for image integration? How can I adjust my prompt to ensure that DALL·E 3 fully incorporates the details of my uploaded PNG image into a bigger new one?

DALL-E is just a tool that ChatGPT can use. It only accepts English text as input which ChatGPT’s AI has written.

Therefore you will not be able to directly upload for the purpose of replicating elements in the image that cannot be solely described.

Images are uploaded for GPT-4o’s vision skill on an image. This can simply be a time-saver versus simply describing everything you want yourself.

Here is an example of understanding the AI’s inability to pass an image to DALL-E 3, but making use of the vision skill:

[mitchell_d00]

One cat is real the others use it as frame with prompt work… if it is a person or other image TOS prevents it…

Did not say it would do it automatically I said it’s a tuff one to do and it needs prompt work for each image uploaded ,anyone can upload an image and say analyze which does not work. It’s hard to do…

[_j]

The output desired is a “bigger new one”. That expresses a desire for a close relationship to the input beyond text.

Often there is misunderstanding, that ChatGPT could do anything at all like “Here’s my picture, show me in a green dress”. It cannot.

[mitchell_d00]

I see yes. That can’t be done. I’m sorry Dalle only does wide narrow and normal but you can play in those ranges. yes it is incredibly hard to reproduce images from other images and make them look even near eachother

[francisco.fhlores]

(post deleted by author)