Stable Diffusion: Refining character details with Inpainting
Stable DiffusionDemo summary
The demo shows a step-by-step refinement of a character's outfit and body structure by masking specific areas and using text prompts to change clothing to 'white battle attire'.
Step-by-step
- Drag the image into the Inpaint tab
- Erase any previous mask paintings
- Highlight the specific sections of the outfit or body you want to modify
- Enter a text prompt for the desired changes, such as 'white battle attire'
- Generate and review the variations for hand and limb structure
- Mask the broken limb or specific areas like shoes to refine them further
Watch out for
- Inpainting can sometimes result in broken limbs, incorrect hand sizes, or poor foot structure that requires further refinement
Tips
- Highlight only specific parts of the image to change the body structure without affecting the general position
- Leave areas like the foot, ankle, or waist unmasked to preserve the overall look of the character while changing the middle section
Highlights
“our first section of images are coming in and they're all looking pretty great”
All demos from “Generate AI Images Without Paying A Dime | Stable Diffusion Tutorial”
- 3:150:33Generate images with FooocusThe user demonstrates entering a text prompt for 'cats on a window lid' into the Fooocus interface and shows the real-time generation process using the local GPU.Fooocus· Text to Image
- 4:000:30Adjusting generation settings in FooocusThe demo shows how to access advanced settings in Fooocus to change performance modes (Speed vs. Quality) and the number of images generated per batch.Fooocus· AI Image Generator
- 5:150:39Prompting for cinematic lightingThe creator demonstrates writing a detailed prompt for a 'cyberpunk city at night' with 'cinematic lighting' and monitors the generation progress in the Fooocus window.Fooocus· Text to Image
- 6:370:41Upscaling images in FooocusThe user demonstrates the 'Image Upscale' feature by dragging a generated city image into the tool and applying a 1.5x upscale to resolve background details.Fooocus· AI Image Upscaler
- 7:350:57Inpainting specific image areasThe demo shows using the brush tool in the Inpaint tab to mask a specific building in an image and regenerate only that section to improve its appearance.Stable Diffusion· AI Inpainting
- 8:350:37Using Improve Detail and Modify ContentThe creator walks through the 'Improve Detail' and 'Modify Content' modes in Fooocus to increase resolution of background elements and change specific objects like cars.Fooocus· AI Inpainting
- 11:070:58Image Prompting with PyraCanny and CPDSThe user demonstrates the 'Image Prompt' feature, explaining how to use PyraCanny to map character positions and CPDS for color/contrast matching from a reference image.Stable Diffusion· Image to Image
- 13:121:10Refining character details with InpaintingCurrentThe demo shows a step-by-step refinement of a character's outfit and body structure by masking specific areas and using text prompts to change clothing to 'white battle attire'.Stable Diffusion· AI Inpainting
- 15:200:30Expanding image canvas (Outpainting)The user demonstrates the 'Content Expansion' tool in Fooocus to convert a portrait image into a landscape format by generating new left and right sections.Fooocus· AI Outpainting
- Watch “Generate AI Images Without Paying A Dime | Stable Diffusion Tutorial” →
AI Inpainting
- 7:350:57Inpainting specific image areasThe demo shows using the brush tool in the Inpaint tab to mask a specific building in an image and regenerate only that section to improve its appearance.Clear
- 13:121:10Refining character details with InpaintingCurrentThe demo shows a step-by-step refinement of a character's outfit and body structure by masking specific areas and using text prompts to change clothing to 'white battle attire'.Clear