Stable Diffusion: Image Prompting with PyraCanny and CPDS
Stable DiffusionDemo summary
The 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.
Step-by-step
- Turn off all advanced features
- Go to the Image Prompt section
- Drag your reference image into the input area
- Scroll to the bottom and click Advanced to reveal the feature bar
- Select either PyraCanny or CPDS based on your visual requirements
Options
- Use PyraCanny to map character positions into the new image
- Use CPDS to match contrast, color, and saturation from the source image
- Use standard Image Prompt to let the AI scan the image as a general suggestion
Watch out for
- The advanced feature bar is hidden by default and requires clicking 'Advanced' at the bottom to access
Tips
- Use PyraCanny if you need to maintain the specific layout or positioning of characters from your reference image
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 CPDSCurrentThe 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 InpaintingThe 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” →
Image to Image
- 11:070:58Image Prompting with PyraCanny and CPDSCurrentThe 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.Clear