stable-diffusion-webui icon indicating copy to clipboard operation
stable-diffusion-webui copied to clipboard

Published 20 hours ago • verified publisher icon AUTOMATIC1111

[Feature Request]: How to get text in the rendered image without being garbled?

Open polavishnu4444 opened this issue 2 years ago • 4 comments

Is there an existing issue for this?

  • [X] I have searched the existing issues and checked the recent builds/commits

What would your feature do ?

Like the GFPGAN for face restoration, is there any alternative for the text to be restored? Even with self-training of the model?

polavishnu4444 avatar Nov 01 '22 01:11 polavishnu4444

you can't restore something that is not there, so no. You'll need photoshop/gimp etc

cmp-nct avatar Nov 01 '22 01:11 cmp-nct

This isn't an issue, at best it's a feature request, but it has nothing to do with the repo, it's an AI thing, nowhere near doing accurate text yet.

BunnyViking avatar Nov 01 '22 02:11 BunnyViking

This is not possible through code. You should close this issue because this is not up to whoever maintaining this repo. It is fully up to the model you are using.

aliencaocao avatar Nov 01 '22 03:11 aliencaocao

In the inpainting flow I was seeing the masked image text as well being garbled. Is that anyone aware with?

polavishnu4444 avatar Nov 01 '22 09:11 polavishnu4444