In-context Region-based Drag: Drag Any Region to Any Shape

2026-06-24

AI Summary

This work addresses the imprecise deformations in point-based drag editing caused by semantic ambiguity by proposing a context-aware, region-level drag editing method. Taking as input a source image along with source and target region masks, the approach leverages a diffusion model enhanced with multimodal attention mechanisms to achieve high-fidelity edits. Its core innovations include two attention regularization strategies: image-mask attention consistency and source-target region attention correspondence. To facilitate training, the authors also introduce PRD, the first large-scale paired-region dataset. Experimental results demonstrate that the proposed method significantly outperforms existing approaches in both quantitative metrics and user studies, achieving superior editing accuracy and visual fidelity.

Abstract

Diffusion models have shown promise in drag-style editing. Previous works mainly focus on point-based drag, which is inherently ambiguous. This paper focuses on region-based drag and introduces a novel In-Context Region-based Drag (ICRDrag) method. Under the in-context learning framework, ICRDrag consumes a source image, a source region mask, and a target region mask, producing the target dragged image. Built upon the basic in-context learning model, we introduce two novel attention regularization: 1) image-mask attention consistency to ensure that a target region attends to similar source regions for image and mask modalities; 2) source-target attention correspondence to ensure the mutual correspondence between source and target regions. To facilitate region-based drag, we also construct Paired Region Dataset (PRD), a large-scale dataset with paired masks and images. Extensive experiments show that ICRDrag significantly outperforms existing methods in both quantitative metrics and user studies, achieving superior editing accuracy and visual fidelity. The dataset, code, and model are available at https://github.com/bcmi/ICRDrag-Region-Drag-Editing.

Problem

• region-based drag
• in-context learning
• diffusion models
• image editing
• attention regularization

Innovation

• region-based drag
• in-context learning
• attention regularization
• diffusion models
• paired region dataset