DALL-E is illiterate with the text it adds in images

What you’re doing is like trying to cut a steak with a spoon and saying that it’s no good because you can’t use it as you expected. This is not a system integration, so your comparison isn’t valid. Take these for what they are, inference models. Please focus on the bold.

Lastly, you don’t need to be a Photoshop expert to fix the images. Even Paint today has AI tools that can help you with that. I’m not sure if you want to address your issue or simply express your frustration, though.

I’ve been trying to use this library:

It could use an update though . . . the problem I have been seeing with using the detextify library is that it doesn’t recognize giberish very well.

So it leaves gibberish and removes text that is actually correct?

In my experience so far, yes. At least it leaves the garbled stuff in place.

I have cleaned it up in many cases with Canva website. They have a tool for adjusting and correcting text in images.

1 Like

Try “FANGO” without the spaces and in quotes… didn’t get 3d here, but might help you…

1 Like

I’ve tried… same results

Yeah, I can’t nudge it either…

DALLE will improve, but in the meantime, I think Adobe has a specific tool for “AI letters” that you can then put together however you like…