ziqihuangg/ReVersion
[SIGGRAPH Asia 2024] ReVersion: Diffusion-Based Relation Inversion from Images
This project helps graphic designers and artists recreate specific visual relationships seen in a few example images. By inputting several images demonstrating a certain interaction (like 'a pattern painted on a surface' or 'an object carved by another'), it learns this unique relationship. You can then use this learned relationship to generate new images where different objects exhibit the same interaction, producing fresh, consistent visuals.
504 stars. No commits in the last 6 months.
Use this if you need to capture a specific visual interaction from a few images and apply it consistently to generate new images with different subjects.
Not ideal if you're looking to generate completely novel images without relying on a pre-defined visual relationship from examples, or if you only need basic image manipulation.
Stars
504
Forks
20
Language
Python
License
—
Category
Last pushed
Oct 07, 2025
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/diffusion/ziqihuangg/ReVersion"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
bytedance/InfiniteYou
🔥 [ICCV 2025 Highlight] InfiniteYou: Flexible Photo Recrafting While Preserving Your Identity
River-Zhang/ICEdit
[NeurIPS 2025] Image editing is worth a single LoRA! 0.1% training data for fantastic image...
muzishen/IMAGHarmony
🧩 IMAGHarmony 🧩: Controllable image editing with consistent object quantity and layout. A...
AMAP-ML/FE2E
[CVPR 2026] Beyond Generation: Advancing Image Editing Priors for Depth and Normal Estimation
TencentARC/BrushNet
[ECCV 2024] The official implementation of paper "BrushNet: A Plug-and-Play Image Inpainting...