Hi all.
On Automatic 1111, I’m struggling to get inpaint upload to work. I’ve heard you need to define your mask first, but I don’t know what that means? And how do you ensure it’s grafted onto the correct part of the image? I can’t see any means of selecting any part of either the base image or the upload image?
Any help is very much appreciated. I am both a noob AND an old person.
Thanks.
If you do not want to do the inpainting directly in the web interface, you’ll have to upload a “alpha mask” of the area that you want to inpaint.
Here’s an example:
Let’s say this is your image. You open it in a editor that supports layers like Gimp. You add a transparent layer on top and draw the area that you want to inpaint over it in black. It should look like this.
Next, you replace the color layer with a white layer. It will look like this:
This is the image that you save and what you upload to the Automatic1111 web interface. Feel free to ask if you have any questions.
It works!
And was surprisingly easy.
Thanks so much for your help. I will.pay it forward!
Awesome, glad I could help!
Now I’ve used it a bit, it seems to me that it doesn’t allow you to insert an image of your choice in that black space; it will just generate what you prompt it to, where directed by your uploaded mask. It therefore doesn’t offer much more functionality than the normal inpainting tool. Is that right? It’s just a more precise method of selection?
Apologies if this was obvious to everyone but me.
Yeah, you got it right. The advantage of this method is that you create masks based on things like outlines and color ranges and at higher resolutions. You can also blur areas of the mask to better blend it into the image.