StyleText: A Large-Scale Dataset and Benchmark for Stylized Scene Text Inpainting
This work addresses the lack of large-scale benchmarks for scene text inpainting with style preservation, providing a standardized evaluation protocol for researchers in document analysis and image editing.
StyleText introduces a large-scale dataset and benchmark for stylized scene text inpainting, comprising 28,518 image-mask-prompt triplets. A FluxFill+LoRA baseline trained on this dataset improves OCR accuracy substantially over initialization while maintaining scene style consistency.
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.