Towards Training-Free Scene Text Editing

2026-03-25

AI Summary

Existing scene text editing methods rely on task-specific training or paired data, limiting their scalability and generalization. This work proposes TextFlow, a framework that achieves, for the first time, fully training-free end-to-end scene text editing. TextFlow integrates Attention Boost (AttnBoost) to guide high-fidelity text rendering and employs Flow Manifold Steering (FMS) to model the visual flow between characters and background, preserving structural and stylistic consistency. The method matches or even surpasses training-dependent approaches in both visual realism and textual accuracy, while supporting plug-and-play editing across multiple languages and diverse scenes. TextFlow demonstrates exceptional cross-domain generalization without any fine-tuning or domain adaptation.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Natural Language Processing: Sentence-level Semantics, Textual Inference, etc.
• Machine Learning: Large Multimodal Models (LMMs)

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Economics, Online Markets and Human Computation: LLM based quality controls for crowd work
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Scene text editing seeks to modify textual content in natural images while maintaining visual realism and semantic consistency. Existing methods often require task-specific training or paired data, limiting their scalability and adaptability. In this paper, we propose TextFlow, a training-free scene text editing framework that integrates the strengths of Attention Boost (AttnBoost) and Flow Manifold Steering (FMS) to enable flexible, high-fidelity text manipulation without additional training. Specifically, FMS preserves the structural and style consistency by modeling the visual flow of characters and background regions, while AttnBoost enhances the rendering of textual content through attention-based guidance. By jointly leveraging these complementary modules, our approach performs end-to-end text editing through semantic alignment and spatial refinement in a plug-and-play manner. Extensive experiments demonstrate that our framework achieves visual quality and text accuracy comparable to or superior to those of training-based counterparts, generalizing well across diverse scenes and languages. This study advances scene text editing toward a more efficient, generalizable, and training-free paradigm. Code is available at https://github.com/lyb18758/TextFlow

Problem

• scene text editing
• visual realism
• semantic consistency
• training-free
• text manipulation

Innovation

• training-free
• scene text editing
• attention boost
• flow manifold steering
• visual realism