Q-Mask：面向OCR的视觉语言模型中基于查询的因果遮罩用于文本锚定
Q-Mask: Query-driven Causal Masks for Text Anchoring in OCR-Oriented Vision-Language Models
AI总结：
本文提出Q-Mask框架，通过因果查询驱动的遮罩解码器提升OCR任务中文本锚定的准确性与稳定性，结合大规模标注数据集提升视觉语言模型对文本区域的识别能力。
摘要：
光学字符识别（OCR）日益被视为现代视觉语言模型（VLMs）的基础能力，使它们不仅能够读取图像中的文本，还能支持现实世界中的视觉问答（VQA）推理。然而，实际应用还需可靠的文本锚点，即准确地将查询文本与其对应的空间区域关联起来。为系统评估这一能力，我们引入TextAnchor-Bench（TABench），一个细粒度文本区域锚定的基准测试，揭示了通用和OCR特定的VLMs仍难以建立准确且稳定的文本锚点。为解决这一限制，我们提出Q-Mask，一种基于因果查询驱动遮罩解码器（CQMD）的精确OCR框架。受链式推理启发，Q-Mask执行因果视觉解码，先按顺序生成查询条件的视觉遮罩，再生成最终OCR输出。这种视觉CoT范式将文本的来源与内容分离，强制在识别前获取 grounded evidence，使推理过程中能够显式构建文本锚点。为训练CQMD，我们构建了TextAnchor-26M，一个大规模图像-文本对标注数据集，包含细粒度的遮罩，对应特定的文本元素，鼓励稳定的文本区域对应关系，并在VLM训练中注入强空间先验。大量实验表明，Q-Mask在多样化的视觉场景中显著提升了文本锚定和理解能力。
英文摘要：
Optical Character Recognition (OCR) is increasingly regarded as a foundational capability for modern vision-language models (VLMs), enabling them not only to read text in images but also to support downstream reasoning in real-world visual question answering (VQA). However, practical applications further require reliable text anchors, i.e., accurately grounding queried text to its corresponding spatial region. To systematically evaluate this capability, we introduce TextAnchor-Bench (TABench), a benchmark for fine-grained text-region grounding, which reveals that both general-purpose and OCR-specific VLMs still struggle to establish accurate and stable text anchors. To address this limitation, we propose Q-Mask, a precise OCR framework built upon a causal query-driven mask decoder (CQMD). Inspired by chain-of-thought reasoning, Q-Mask performs causal visual decoding that sequentially generates query-conditioned visual masks before producing the final OCR output. This visual CoT paradigm disentangles where the text is from what the text is, enforcing grounded evidence acquisition prior to recognition and enabling explicit text anchor construction during inference. To train CQMD, we construct TextAnchor-26M, a large-scale dataset of image-text pairs annotated with fine-grained masks corresponding to specific textual elements, encouraging stable text-region correspondences and injecting strong spatial priors into VLM training. Extensive experiments demonstrate that Q-Mask substantially improves text anchoring and understanding across diverse visual scenes.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
