Feature Req: Multiple AI Model with Text Detection

Ability to choose (for example) “Art & CG” AI model, with text detection. The parts with text will have a mask automatically applied, to which you can modify with a variety of different tools (i.e. brush painting, circle/box selection). You can then use a gaussian blur to fade the mask smoothly. The mask will enable you to use a second AI model (i.e. Text & Shapes)

This would make it convenient to Ai Upscale scans of comic books. Comic Book art looks great when using the Art & CG ai model, but the text requires a different AI Model. This is why I think it’s important to be able to use multiple AI models.

Also, I noticed that the Text & Shapes Ai model struggles with text that is slightly rotated. This can happen if the page was scanned poorly with text slanted due to the page not being perfectly flat with the scanner. the AI model should really be able to take into account text that has been rotated, either on purpose or on accident.

Thank you.

Awesome idea. Thank you for sharing.