提出新基准与框架，显著提升复杂表格的解析准确率。
From Diagnosis to Correction: Benchmarking and Improving Real-World Table Parsing
- 分步式框架DEC：拆解-增强-修正，利用视觉一致性引导纠错
- 在真实表格数据集上，结构错误减少2.62点，大表解析提升5.66点
- 无需重训练，适用于现有冻结模型，适合工业级文档处理场景
近期文档解析器在OmniDocBench v1.6上的表格TEDS得分超过93，但社区反馈和审计显示其在复杂真实表格上仍存在持续失败。为量化这一差距，我们引入TableParseMap，一个包含916张真实表格的诊断基准，涵盖五类挑战场景与九种故障类型。最强评估解析器仅达85.03 TEDS，表明整体得分掩盖了显著弱点。分析指出三大互补缺陷：大表格超出单次处理可靠范围，弱或模糊视觉线索阻碍结构感知，重构表格与图像视觉不一致。为此提出DEC（Decompose--Enhance--Correct）——一种视觉一致性引导的代理框架，无需重训练即可提升冻结解析器性能。DEC使用通用视觉语言模型（VLM）作为控制器：分解沿结构感知边界拆分大表，增强暴露弱视觉证据并重解析变换视图，纠正诊断并修复残余错误。视觉一致性门控（VC-Gate）选择性触发干预，视觉一致性排序器（VC-Ranker）验证候选更新并支持回滚，无需推理时的真值HTML。我们进一步通过离线指标与跨模型共识，从4,556个候选中构建1,977表的共识难例集。在三个冻结解析器上，DEC平均提升TEDS 1.57点；在TableParseMap上，总体提升达1.89点，结构错误改善2.62点，大表解析提升5.66点。
原文摘要 · Abstract (English)
Recent document parsers achieve table TEDS scores above 93 on OmniDocBench v1.6, yet community feedback and our audit reveal persistent failures on complex real-world tables. To quantify this gap, we introduce TableParseMap, a diagnostic benchmark of 916 real-world tables organized into five challenging scenarios and nine failure types. The strongest evaluated parser achieves only 85.03 TEDS, showing that aggregate benchmark scores conceal substantial weaknesses. Our analysis attributes these failures to three complementary limitations: large tables exceed the reliable processing scale of a single pass, weak or ambiguous visual cues hinder structure perception, and the reconstructed table may remain visually inconsistent with the image. We therefore propose DEC (Decompose--Enhance--Correct), a visual-consistency-guided agentic framework that improves frozen table parsers without retraining. DEC uses a general VLM as the controller: Decompose partitions large tables along structure-aware boundaries, Enhance exposes weak visual evidence and reparses transformed views, and Correct diagnoses and repairs residual errors. A Visual Consistency Gate (VC-Gate) selectively triggers intervention, while a Visual Consistency Ranker (VC-Ranker) verifies candidate updates and supports rollback without ground-truth HTML at inference time. We further derive a 1,977-table Consensus-Hard Set from 4,556 candidates through offline metrics and cross-model consensus. Across three frozen parsers, DEC improves TEDS by 1.57 points on average; on TableParseMap, gains reach 1.89 points overall, 2.62 on structural errors, and 5.66 on large tables.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。