StyleText: A Large-Scale Dataset and Benchmark for Stylized Scene Text Inpainting

TL;DR

StyleText introduces a comprehensive dataset and benchmark for evaluating stylized scene text inpainting, emphasizing style preservation and text legibility through controlled, large-scale data and standardized metrics.

Contribution

The paper provides a novel large-scale dataset, an automated construction pipeline, and a reproducible evaluation protocol for scene text inpainting with style preservation.

Findings

FluxFill+LoRA baseline improves OCR accuracy significantly.

The dataset enables controlled evaluation of style and text legibility.

Reproducible metrics facilitate consistent benchmarking.

Abstract

We present StyleText, a large-scale dataset and benchmark for localized scene-text inpainting with style preservation. StyleText contains 28,518 image-mask-prompt triplets grouped into 9,932 scene families, enabling controlled evaluation of text legibility and visual consistency under shared scene context. We construct the dataset with an automated pipeline that combines LLM prompt templating, Flux-based source generation with key-value (KV) cache injection, OCR-based semantic filtering, polygon mask extraction, and mask-conditioned FluxFill augmentation. We define a reproducible evaluation protocol using normalized OCR metrics (word accuracy and character error rate) and CLIP image-image similarity with explicit preprocessing. A FluxFill+LoRA baseline trained on StyleText improves OCR accuracy substantially over initialization while maintaining scene style consistency, establishing a strong reference point for future comparisons.