Paper page - Real In, Real Out: What If We Only Use Real Data for Scene Text Editing?

Source: https://huggingface.co/papers/2610.27592

Abstract

TextRIRO enables scene text editing using self-paired real data via image disruptions and a conditional diffusion model, improving realism and supporting downstream recognition.

Aiming to modify textual content while preserving the original style,Scene Text Editing(STE) has significant practical value in many applications. STE typically requires paired images before and after editing for training. However, recent approaches rely heavily on synthetic paired data and complex disentanglement strategies. They struggle to faithfully capture the diversity and complexity of natural scenes, often producing artifacts or hallucinations. In this work, we challenge the common practice and investigate whether STE can be driven by turning each real sample into its own paired supervision. Concretely, we explore multiple ways to disrupt the original image (e.g., Crop, Shuffle, Mask) to constructself-paired trainingdata, and build a strong baseline (TextRIRO}) upon aconditional diffusionarchitecture. Our experiments show that combining character-column shuffling with image cropping (Shuffle&Crop) effectively destroys text semantics while preserving key style cues, enabling the model to learn editing behaviors directly from the real text image. With this concise yet real-oriented design,TextRIROachieves promising performance in both accuracy and visual realism across real-world benchmarks, substantially mitigating thesynthetic-to-real domain gap. Furthermore, we leverageTextRIROto buildTextRIRO-3M, a large-scale realistic dataset forlong-text recognitionby concatenating edited and original images. Experimental results show that training recognition models withTextRIRO-3M significantly improves recognition accuracy, demonstrating thatTextRIROprovides highly useful training resources for downstream tasks. Overall,TextRIROlifts STE to realistic, reliable, and widely applicable new levels. Code is available at https://github.com/YesianRohn/TextRIRO.

View arXiv pageView PDFProject pageGitHub1Add to collection

Get this paper in your agent:

hf papers read 2610\.27592

Don’t have the latest CLI?curl \-LsSf https://hf\.co/cli/install\.sh \| bash

Models citing this paper1

#### Yesianrohn/TextRIRO Image-to-Image• Updatedabout 13 hours ago

Datasets citing this paper2

#### Yesianrohn/RIROCommon Updatedabout 13 hours ago • 83 #### Yesianrohn/TextRIRO-3M Updatedabout 13 hours ago • 35

Spaces citing this paper1

Collections including this paper0

No Collection including this paper

Add this paper to acollectionto link it from this page.