What it does?
Since Stable Diffusion was released, there's been lots of tools coming out that allow you to generate AI images. I was messing around with finetuning Stable Diffusion to try and get it to generate images of myself but most of the outputs needed some additional editing. I started this project because I figured I could create something to edit these pictures - though I eventually moved away from editing AI generated images (there are lots of tools that can do this) and focused on editing real images which is more difficult because real images tend to have a much higher resolution.
The application itself had a number of features:
- Uploading images from local storage
- Generating images with stable diffusion
- Create image masks using brush tool to click on where you want the mask to be
- Generate masks for objects in the image using Segment Anything
- Inpaint these masks
- User configurable prompts, inference steps and guidance
I stopped working on this project when Photoshop released most of these features in their beta product, though mine was quite far advanced.