Skip to content

[ICLR 2026] The official code for "Geometric Image Editing via Effects-Sensitive In-Context Inpainting with Diffusion Transformers"

License

Notifications You must be signed in to change notification settings

PRIS-CV/GeoEdit

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

24 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

🗝 GeoEdit

Project Page arXiv

Code release for "Geometric Image Editing via Effects-Sensitive In-Context Inpainting with Diffusion Transformers"

Abstract: Recent advances in diffusion models have significantly improved image editing. However, challenges persist in handling geometric transformations, such as translation, rotation, and scaling, particularly in complex scenes. Existing approaches suffer from two main limitations: (1) difficulty in achieving accurate geometric editing of object translation, rotation, and scaling; (2) inadequate modeling of intricate lighting and shadow effects, leading to unrealistic results. To address these issues, we propose GeoEdit, a framework that leverages in-context generation through a diffusion transformer module, which integrates geometric transformations for precise object edits. Moreover, we introduce Effects-Sensitive Attention, which enhances the modeling of intricate lighting and shadow effects for improved realism. To further support training, we construct RS-Objects, a large-scale geometric editing dataset containing over 120,000 high-quality image pairs, enabling the model to learn precise geometric editing while generating realistic lighting and shadows. Extensive experiments on public benchmarks demonstrate that GeoEdit consistently outperforms state-of-the-art methods in terms of visual quality, geometric accuracy, and realism.

News

  • Feb 9 2026: 🔥 Our Project Page has been published!
  • Jan 26 2026: 🔥 🔥 GeoEdit has been accepted to ICLR'26!
TODO Task Update
📂 Dataset Under active preparation
⚖️ Model Weights Under active preparation

🛠️ Installation

conda create -n geoedit python==3.10

conda activate geoedit

pip install -r requirements.txt

🚀 Training

  • Update FLUX model paths: In experiments/experiments.yaml, replace /path/to/black-forest-labs-FLUX.1-Fill-dev and /path/to/black-forest-labs-FLUX.1-Redux-dev with your actual paths.

  • Run the training script: Execute the training using the command:

    bash scripts/train.sh

About

[ICLR 2026] The official code for "Geometric Image Editing via Effects-Sensitive In-Context Inpainting with Diffusion Transformers"

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published