以由粗到细的视觉处理提升文档解析效率与性能
Boosting Document Parsing Efficiency and Performance with Coarse-to-Fine Visual Processing
VRFM 用定位与上下文关系预测筛出有效视觉 token、抑制背景冗余，这一「先粗筛区域再细识别」的 token 选择思路可直接迁移到高分辨率图像输入下的任意 VLM 任务（如高分辨率 OCR、图表理解、遥感图文），在保持精度的同时大幅削减视觉 token 数与推理成本。
原文依据 a lightweight Valid Region Focus Module (VRFM) which leverages localization and contextual relationship prediction capabilities to identify valid vision tokens
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论文档解析是一项细粒度任务，图像分辨率会显著影响其性能。尽管基于视觉语言模型的先进研究受益于高分辨率输入以提升模型性能，但这往往导致视觉 token 数量呈二次增长，并显著抬高计算成本。我们将这种低效归因于文档图像中存在大量视觉区域冗余，例如背景。为解决该问题，我们提出 PaddleOCR-VL，一种新颖的由粗到细架构，聚焦于语义相关区域并抑制冗余区域，从而同时提升效率与性能。具体而言，我们引入轻量级的有效区域聚焦模块（VRFM），它利用定位与上下文关系预测能力来识别有效视觉 token。随后，我们设计并训练了一个紧凑而强大的 0.9B 视觉语言模型（PaddleOCR-VL-0.9B）来执行精细识别，并以 VRFM 的输出为引导，避免直接处理整幅大图。大量实验表明，PaddleOCR-VL 在页面级解析与元素级识别上均达到最优性能。它显著优于现有方案，与顶级视觉语言模型相比展现出很强的竞争力，并在使用大幅更少的视觉 token 与参数的同时实现快速推理，凸显了有针对性的由粗到细解析对于准确且高效的文档理解的有效性。源代码与模型已公开于 https://github.com/PaddlePaddle/PaddleOCR。
英文原文摘要 1615 字符
Document parsing is a fine-grained task where image resolution significantly impacts performance. While advanced research leveraging vision-language models benefits from high-resolution input to boost model performance, this often leads to a quadratic increase in the number of vision tokens and significantly raises computational costs. We attribute this inefficiency to substantial visual regions redundancy in document images, like background. To tackle this, we propose PaddleOCR-VL, a novel coarse-to-fine architecture that focuses on semantically relevant regions while suppressing redundant ones, thereby improving both efficiency and performance. Specifically, we introduce a lightweight Valid Region Focus Module (VRFM) which leverages localization and contextual relationship prediction capabilities to identify valid vision tokens. Subsequently, we design and train a compact yet powerful 0.9B vision-language model (PaddleOCR-VL-0.9B) to perform detailed recognition, guided by VRFM outputs to avoid direct processing of the entire large image. Extensive experiments demonstrate that PaddleOCR-VL achieves state-of-the-art performance in both page-level parsing and element-level recognition. It significantly outperforms existing solutions, exhibits strong competitiveness against top-tier VLMs, and delivers fast inference while utilizing substantially fewer vision tokens and parameters, highlighting the effectiveness of targeted coarse-to-fine parsing for accurate and efficient document understanding. The source code and models are publicly available at https://github.com/PaddlePaddle/PaddleOCR.