LexiAlign: A Diffusion Model Text Alignment and Refinement Method Based on Local Regeneration

Despite the recent advances of diffusion models in high-quality image synthesis, generating visually accurate textual content remains a persistent bottleneck—commonly manifesting as misspellings, glyph distortions, and semantic drift. We introduce \textbf{LexiAlign}, a \emph{language-guided local diffusion refinement} framework that directly targets these issues through three tightly coupled components: robust \emph{optical character recognition} (OCR)-based text extraction, \emph{language model}-driven semantic correction, and high-fidelity local inpainting via masked diffusion. Unlike prior approaches that retrain large diffusion backbones or overwrite entire regions, LexiAlign performs \emph{character-level targeted repair} while preserving surrounding visual context and style. To support systematic training and evaluation, we construct \textbf{SynOCRText}, a 120k-sample benchmark covering 8 languages, over 20 fonts, diverse layouts, and fine-grained error masks. On SynOCRText, LexiAlign achieves \textbf{88.4\% OCR accuracy} (+6.3\% over the best baseline), Contrastive Language-Image Pretraining (CLIP) Score of 0.852 (+0.023), Peak Signal-to-Noise Ratio (PSNR) of 30.92\,dB (+2.51\,dB), and Structural Similarity Index Measure (SSIM) of 0.893 (+0.020). These results establish LexiAlign as a \emph{plug-and-play, domain-agnostic} solution for reliable visual text alignment, offering both \emph{quantitative superiority} and \emph{practical deployability} for creative design, advertising, and multimodal content generation.