Inpainting: The Digital Surgeon's Scalpel
Inpainting is one of the most powerful workflows in Generative AI. It allows you to modify specific parts of an image while preserving the overall composition. It is the bridge between random generation and controlled art direction.
1. The Mask: Defining Constraints
The mask is a binary map (usually black and white). White areas are "allowed to change," while black areas are "frozen." A soft brush edge helps blend the new generation into the existing pixels seamlessly.
2. Denoising Strength
This is your "Creative Freedom" slider.
Low Strength (0.1 - 0.3)
Retains the original shape and color. Good for changing materials or removing noise without changing geometry.
High Strength (0.7 - 1.0)
Destroys original pixels. Allows for generating completely new objects that weren't there before.
3. Inpaint at Full Resolution
Often called "Inpaint Only Masked". The AI focuses all its processing power on just the masked area, upscaling it, generating detail, and then shrinking it back down to fit. This creates incredible detail for faces and hands.
Pro Tip: Good inpainting is 80% masking accuracy and 20% prompting.