AI_SDXL_inpainting_with_GroundedSAM2 - 100-hours-a-week/16-Hot6-wiki GitHub Wiki

μƒμœ„ λ¬Έμ„œλ‘œ 이동 : AI Wiki

AI_SDXL_inpainting_with_GroundedSAM2

  • The SDXL Inpainting model edits or fills parts of an image based on a mask and a text prompt.


Model Test

  • Origin Image test3

  • Mask Image inpainting_background_mask We used Grounded SAM2 for masking to precisely define the areas for SDXL Inpainting. Grouded with SAM2

Example Output

1 2 3
inpainted_result-2 inpainted_result-3 inpainted_result-4
  • These images were generated using different prompts to explore various outputs.

Issue

  • The mask images are a bit jagged, which results in less clean inpainting outputs.
  • Imprecise masking of the 'desk' and 'deskmat' appears to cause issues in the final output.

Next Steps

  • Generate a mask image that excludes the desk and deskmat from the inpainting area.

Reference

diffusers/stable-diffusion-xl-1.0-inpainting-0.1

⚠️ **GitHub.com Fallback** ⚠️