Stable Diffusion Masterclass - Theory, Code & Application

Masking - Image to Image Generation

Now that we understand how image to image generation works, we'll explore a couple of other cool tools that we can use to have better control over the generated images, starting with masking.

When generating images using an input image, masking is a technique that allows us to specify which parts of the input image we want to influence the output image. We can use masking to specify which parts of the input image we want to keep and which parts we want to remove or rewrite.