DataEvolver: 面向文本丰富图像生成的自我进化多智能体数据构建

摘要

文本丰富图像生成是图像生成中最具挑战性的场景之一，因为模型必须同时生成视觉上逼真的图像，并呈现清晰可读、语义对齐且布局一致的文本。现有数据流水线通常遵循静态的"爬取-过滤-冻结"范式：先收集候选样本，一次性过滤，再固定采用通过的数据用于训练。然而，被拒绝的样本往往被丢弃，尽管它们通常包含有用的失败信号（如OCR错误和语义不匹配），导致后续构建轮次可能重复相同的失败模式。为解决这些局限，我们提出DataEvolver——一个用于文本丰富图像数据构建的自演化多智能体框架。DataEvolver将数据构建视为反馈驱动的构建策略演化过程：检索器收集候选样本，验证器分配质量分数和拒绝原因，评论家将轮次级反馈总结为语义反馈，生成器通过定向合成覆盖低覆盖区域。更新后的反馈记忆随后指导下一轮构建。在文本丰富图像生成基准上的实验表明，在匹配数据预算的情况下，DataEvolver比固定数据集基线生成更有用的训练数据。在PixArt-alpha的0.75M规模上，DataEvolver在TextScenesHQ上比最强基线提升了85.3%的OCR-F1，在LongTextBench上提升了35.3%。这些改进在两个评估基准上保持一致，并迁移至Show-o2，表明DataEvolver的收益不依赖于单一的下游生成器。这些结果表明，被拒绝的样本可为改进文本丰富图像数据构建提供可操作的反馈。

English

Text-rich image generation is one of the most challenging settings in image generation, since models must simultaneously produce visually realistic images and render legible, semantically aligned, and layout-consistent text. Existing data pipelines usually follow a static crawl-filter-freeze paradigm. They collect candidate samples, filter them once, and freeze the accepted data for training. However, rejected samples are usually discarded, although they often contain useful failure signals such as OCR errors and semantic mismatches. As a result, later construction rounds may repeat the same failure modes. To address these limitations, we propose DataEvolver, a self-evolving multi-agent framework for text-rich image data construction. DataEvolver treats data construction as feedback-driven construction policy evolution. A Retriever collects candidate samples, a Verifier assigns quality scores and rejection causes, a Critic summarizes round-level feedback into semantic feedback, and a Generator completes under-covered regions through targeted synthesis. The updated feedback memory then guides the next construction round. Experiments on text-rich image generation benchmarks show that DataEvolver produces more useful training data than fixed-dataset baselines under matched data budgets. At the 0.75M scale on PixArt-alpha, DataEvolver improves OCR-F1 over the strongest baseline by 85.3 percent on TextScenesHQ and 35.3 percent on LongTextBench. The improvements are consistent across both evaluated benchmarks and also transfer to Show-o2, indicating that the benefit of DataEvolver is not tied to a single downstream generator. These results suggest that rejected samples can provide actionable feedback for improving text-rich image data construction.