Training-free Geometric Image Editing on Diffusion Models

2025-07-31

AI Summary

This work addresses the challenge of geometric image editing in diffusion models—specifically, achieving precise object localization, orientation control, and non-rigid deformation while preserving global scene coherence. We propose a training-free, disentangled editing framework that decomposes editing into three sequential stages: geometric transformation of the target object, source-region inpainting, and target-region detail refinement. Leveraging FreeFine, our approach enables high-fidelity, zero-shot 2D/3D geometric editing without model fine-tuning. Our key innovation lies in the explicit decoupling of geometric operations from diffusion-based reconstruction—a first in the literature—thereby eliminating distortions and semantic drift inherent in end-to-end fine-tuning. Evaluated on our newly introduced GeoBench benchmark, the method achieves state-of-the-art performance in large-scale deformations and complex structural edits, demonstrating superior image fidelity and geometric accuracy compared to existing approaches.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning
• Machine Learning: Deep Generative Models & Autoencoders

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Security and Privacy: Data transparency and provenance

Abstract

We tackle the task of geometric image editing, where an object within an image is repositioned, reoriented, or reshaped while preserving overall scene coherence. Previous diffusion-based editing methods often attempt to handle all relevant subtasks in a single step, proving difficult when transformations become large or structurally complex. We address this by proposing a decoupled pipeline that separates object transformation, source region inpainting, and target region refinement. Both inpainting and refinement are implemented using a training-free diffusion approach, FreeFine. In experiments on our new GeoBench benchmark, which contains both 2D and 3D editing scenarios, FreeFine outperforms state-of-the-art alternatives in image fidelity, and edit precision, especially under demanding transformations. Code and benchmark are available at: https://github.com/CIawevy/FreeFine

Problem

• Repositioning, reorienting, or reshaping objects in images
• Preserving scene coherence during geometric image editing
• Handling large or complex transformations effectively

Innovation

• Decoupled pipeline for geometric image editing
• Training-free diffusion approach FreeFine
• Outperforms in fidelity and precision