InnoText: A Unified Model for Visual Text Generation and Editing

2026-07-24

AI Summary

Existing methods for visual text generation and editing suffer from limitations in structural regularity, readability—particularly for small fonts and Chinese characters—and consistency across diverse tasks. This work proposes the first unified Diffusion Transformer (DiT) architecture that jointly models both generation and editing within a single framework. By introducing font-size-aware modulation (FSAM), a small-character enhancement strategy, and a task-region-weighted loss function, the approach significantly improves cross-task generalization and fine-grained text fidelity. Evaluated on a newly curated bilingual (Chinese–English) visual text dataset, the method consistently outperforms current state-of-the-art techniques across various fonts, font sizes, and complex backgrounds, achieving leading performance in both textual accuracy and visual realism.

Abstract

Diffusion models have recently achieved remarkable success in high-fidelity image synthesis, yet their application to visual text generation and editing remains relatively underexplored. Unlike general image generation, visual text tasks demand precise structural regularity and legibility, which may pose additional challenges for small-scale text and non-Latin scripts such as Chinese. Existing UNet-based models often struggle to produce clear and coherent text, while DiT-based models, though more expressive, are typically limited to a single task, which may lead to redundant training pipelines, inconsistent visual styles, and reduced cross-task generalization. To address these challenges, we propose InnoText, a unified DiT-based framework capable of performing both text generation and editing within a single model. We introduce a Font Size-Aware Modulation (FSAM) module to enhance representations across font scales, a Small-Character Aware Augmentation strategy to improve fine-grained fidelity, and a Task-Specific Region Weighted Loss for adaptive optimization. To support training and evaluation, we also construct a high-quality bilingual (English-Chinese) visual text dataset covering diverse fonts, sizes, and backgrounds. Experimental results demonstrate that our method achieves superior generation accuracy and editing quality, producing visually appealing and realistic text images.

Problem

• visual text generation
• text editing
• diffusion models
• non-Latin scripts
• cross-task generalization

Innovation

• unified diffusion model
• visual text generation
• text editing
• font size-aware modulation
• small-character augmentation