https://ym-xu.github.io/living-survey/fields/document-understanding/#tag=parsingTurning pages into structured text: PDF-to-Markdown, formulas, end-to-end parsers.Thu, 12 Nov 2026 15:12:01 +0000https://arxiv.org/abs/2611.10679https://arxiv.org/abs/2611.10679Sat, 07 Nov 2026 00:00:00 +0000Halimat Afolabi, Daeyoung Lee Document parsing benchmarks are often treated as stable measures of model quality, yet rankings may depend substantially on benchmark design choices. In this work, we present RankShift, a systematic study comparing OCR and vision-language models (VLMs) under controlled variation in dataset composition, matching strateg…https://arxiv.org/abs/2611.06449https://arxiv.org/abs/2611.06449Wed, 04 Nov 2026 00:00:00 +0000Boyu Niu, Yingjie Xiong, Shurui Li, Linhan Pang, et al. Traditional document parsing systems use Markdown as their output representation, which is effective for linear text but does not fully preserve fine-grained spatial structure, table geometry, visual hierarchy, or rendering semantics. We show that HTML is advantageous for heterogeneous and layout-sensitive documents wh…https://arxiv.org/abs/2610.39308https://arxiv.org/abs/2610.39308Thu, 29 Oct 2026 00:00:00 +0000Shu Wei, Jingjing Wu, Qunyi Xie, Lingshu Zhang, et al. Unlimited OCR ushers the era of long-horizon parsing, substantially mitigating the growth of working memory as document lengthens. However, as document length scales to hundreds of pages, the visual context grows substantially, sharply increasing the computational cost of each decoding step and limiting the scalability…https://arxiv.org/abs/2611.00092https://arxiv.org/abs/2611.00092Thu, 22 Oct 2026 00:00:00 +0000Aman Rangapur, Antonio Jimeno Yepes, Daniel Bustamante, Nutchapol Dendumrongsup, et al. We introduce OCRia-1, a 4B vision-language model (VLM) for document parsing that excels at preserving table structure, extracting chart data and retaining text formatting. OCRia-1 handles content that has traditionally been challenging in document parsing: merged cells and spanning headers, tables with thousands of cel…https://arxiv.org/abs/2610.29188https://arxiv.org/abs/2610.29188Wed, 21 Oct 2026 00:00:00 +0000Eunbin Kim, Hyeonseok Kang, Sangkeun Jung · EMNLP 2026 Tables in real-world documents, including scientific papers, financial reports, and statistical summaries, frequently use merged cells, multi-level headers, and sectioned layouts. Yet the lightweight formats LLMs handle best (Markdown, CSV) cannot express these structures, while expressive formats (HTML, LaTeX) impose …https://arxiv.org/abs/2610.25235https://arxiv.org/abs/2610.25235Sun, 18 Oct 2026 00:00:00 +0000Jumana Alsubhi, Ahmed Alhusayni, Abdulrahman Gharawi, Israa Hamdine, et al. Choosing an OCR system for a multilingual archive requires balancing accuracy against deployment constraints. This study compares cloud and self-hosted systems on three tasks: bounding box detection on 48 Arabic, English, and French pages, text recognition on 315 pages across nine document types, and mathematical expre…https://arxiv.org/abs/2610.21979https://arxiv.org/abs/2610.21979Thu, 15 Oct 2026 00:00:00 +0000Yong Zu, Ninghua Yang, Zezhong Tan, Xinhong Ma, et al. End-to-end full-page document parsing must preserve the content and structure of heterogeneous elements, yet otherwise accurate outputs can contain local errors and omissions. A single page-level advantage does not explicitly identify which elements differ in quality across sampled outputs or where their errors occur. …https://arxiv.org/abs/2610.20335https://arxiv.org/abs/2610.20335Wed, 14 Oct 2026 00:00:00 +0000Yuheng Chen, Yufan Chen, Zhuojun Cai, Ruiping Liu, et al. · NeurIPS 2026 Document understanding systems are typically evaluated as separate components via layout analysis, optical text recognition, reading-order prediction, and image captioning. This separation leaves scientific figures, captions, and visual semantics weakly connected, and fails to measure whether a model can produce a sing…https://arxiv.org/abs/2610.12363https://arxiv.org/abs/2610.12363Thu, 08 Oct 2026 00:00:00 +0000Xiazhen Wu, Wansong Qin, Yangbin Zheng, Liangda Fang, et al. Intelligent grading and automated scoring technologies constitute critical infrastructure for smart education. However, existing document parsing and handwriting recognition benchmarks are predominantly designed for well-structured printed documents or isolated mathematical expressions, lacking datasets that capture th…https://arxiv.org/abs/2609.36136https://arxiv.org/abs/2609.36136Mon, 28 Sep 2026 00:00:00 +0000Xin Chen, Anan Du, Feng Feng, Pei Fu, et al. Compact OCR-specific vision-language models achieve strong document parsing performance, but often rely on costly supervision and focus primarily on visual-text reconstruction. We introduce Xiaomi-OCR-0, a unified 0.8B model for document parsing and OCR-centric understanding. We build an approximately 170M-sample OCR-c…https://arxiv.org/abs/2609.34841https://arxiv.org/abs/2609.34841Mon, 28 Sep 2026 00:00:00 +0000Zixiao Dong, Wei Yang, Zihao Liu, Chenshu Li, et al. A well-designed extraction schema is not necessarily ready for reliable LLM execution. When only limited verified extractions are available, manually tuning hundreds of field definitions through trial and error is costly. We frame this problem as few-instance schema calibration: adapting the operational semantics of an…https://arxiv.org/abs/2609.32513https://arxiv.org/abs/2609.32513Sat, 26 Sep 2026 00:00:00 +0000Paulo Grane Gabriel Silva, Lorenz Bernard Marqueses, Joel Ilao Handwritten mathematical expression recognition (HMER) refers to the task of recognizing and converting handwritten mathematics into a parsable markup language, usually LaTeX. No current state-of-the-art-competitive system adjusts to the way a specific person writes, and the domain gap between training images (usually …https://arxiv.org/abs/2609.24220https://arxiv.org/abs/2609.24220Mon, 21 Sep 2026 00:00:00 +0000Uday Allu, Abhivanth Sivaprakash, Pratik Singh, Aman Manocha Retrieval-Augmented Generation (RAG) systems over enterprise knowledge bases must ingest heterogeneous document formats -- PDFs, Word documents, presentations, and scans -- whose content is locked inside complex visual layouts, multi-column pages, and dense tables. Rule-based extraction and OCR destroy reading order, f…https://arxiv.org/abs/2609.23592https://arxiv.org/abs/2609.23592Sun, 20 Sep 2026 00:00:00 +0000Xingyu Lin, Dehui Du End-to-end document parsers increasingly offer an optional reasoning mode for complex pages. On a 180-page entropy-stratified discovery sample with one frozen 4B checkpoint, complexity is the wrong decision variable. Reasoning lowers mean quality by 2.21 Overall at 1.54x tokens; a preregistered input-only model cannot …https://arxiv.org/abs/2609.20423https://arxiv.org/abs/2609.20423Thu, 17 Sep 2026 00:00:00 +0000Hao Yu, Kang Liu, Linnan Zhao, Jiabo Zhan, et al. Document parsing converts document images into structured content and requires reliable performance across diverse layouts and acquisition conditions. Yet training corpora are biased toward common document types and clean digital pages, while expanding coverage alone does not specify how to address a parser's remaining…https://arxiv.org/abs/2609.12917https://arxiv.org/abs/2609.12917Fri, 11 Sep 2026 00:00:00 +0000Shiliang Luo Handwritten mathematical expression recognition (HMER) is conventionally scored by exact-match rates and string-similarity metrics that are blind to where an error occurs: two predictions with identical token-error counts receive identical scores whether they misplace a subscript or swap the operands of a fraction. Ren…https://arxiv.org/abs/2609.03181https://arxiv.org/abs/2609.03181Wed, 02 Sep 2026 00:00:00 +0000Alejandro Barón García, Feng Wang, Emilia Garcia Casademont, Han Xiao We present Jina-OCR-v1, an end-to-end document parsing model built to serve on low-budget GPUs. It combines the compressed-vision encoder and the 3B mixture-of-experts decoder of DeepSeek-OCR, which activates about 570M parameters per token, with a FastMTP speculative decoding head that shares a single draft block recu…https://arxiv.org/abs/2608.29477https://arxiv.org/abs/2608.29477Sun, 30 Aug 2026 00:00:00 +0000Jason Luo, Saibilila Abudukelimu, Judy Song, Andrew Feng, et al. Document question-answering systems increasingly answer questions over collections of retrieved documents rather than one clean source, so robustness to distracting context matters as much as reading ability. When such systems fail, it is often unclear whether the context was too long or the distractors were too close …https://arxiv.org/abs/2608.28698https://arxiv.org/abs/2608.28698Thu, 27 Aug 2026 00:00:00 +0000Mingxu Chai, Chenyu Liu, Ziyu Shen, Jiazheng Zhang, et al. Compared with typical vision-language tasks, document parsing places stronger demands on fine-grained visual perception. Existing vision-language model (VLM)-based parsing approaches rely on globally compressed visual tokens, where fine-grained details are entangled within a single representation and repeatedly accesse…https://arxiv.org/abs/2608.22842https://arxiv.org/abs/2608.22842Mon, 24 Aug 2026 00:00:00 +0000Hang Wang, Jin Zhang, Guoliang Xu, Pengyue Lu, et al. Financial document parsing requires accuracy, structural consistency, and verifiability that current benchmarks often fail to reflect. We present FinixDoc, an end-to-end agentic parsing system for real-world financial documents, with FinixDoc-VL, a 4B-scale vision-language model built on Qwen3-VL-4B, as its core parser…https://arxiv.org/abs/2608.17550https://arxiv.org/abs/2608.17550Tue, 18 Aug 2026 00:00:00 +0000Rihui Jin, Jun Wang, chengyuan zhu, Liang Mingyu, et al. · ACM MM 2026 Academic papers are a primary carrier of scientific knowledge, yet most of this knowledge remains locked in PDFs that are optimized for human reading rather than machine use. For Multimodal Large Language Models (MLLMs), the core challenge is not only perception, but representation: scientific pages interleave text wit…https://arxiv.org/abs/2608.12898https://arxiv.org/abs/2608.12898Thu, 13 Aug 2026 00:00:00 +0000Peng Cai, Zhaofan Zou, Shifa Liu, Yikun Wang, et al. Document parsing aims to transform unstructured documents into structured and machine-readable representations. Recent advances in Vision-Language Models (VLMs) have significantly advanced document parsing. However, existing approaches still face two major challenges. First, decoupled VLM-based methods heavily rely on …https://arxiv.org/abs/2608.06758https://arxiv.org/abs/2608.06758Fri, 07 Aug 2026 00:00:00 +0000Shi Chen, Hayato Aida, Makoto Morinaga, Shohei Tanaka, et al. We present Stockmark-Nemotron-3-Nano-Omni-JapanDocReader, a Japanese document understanding model built from Nemotron-3-Nano-Omni-30B-A3B-Reasoning-BF16. The central goal of this work is structured document parsing via capability injection and forgetting control: we inject Japanese structured document parsing capabilit…https://arxiv.org/abs/2608.06146https://arxiv.org/abs/2608.06146Thu, 06 Aug 2026 00:00:00 +0000Hao Yu, Jiabo Zhan, Kang Liu, Linnan Zhao, et al. End-to-end document parsers provide a unified interface, but serialize page layouts and regional contents into one autoregressive sequence. This formulation forces independent regions onto a decoding path whose length grows with the total content, whereas crop-based two-stage parsers expose region-level parallelism at …https://arxiv.org/abs/2608.03525https://arxiv.org/abs/2608.03525Tue, 04 Aug 2026 00:00:00 +0000Haote Yang, Jiang Wu, Jingchao Wang, Xingjian Wei, et al. In organic chemistry papers and patents, molecular structures, reaction schemes, and experimental conditions are often presented as molecular structure depictions, reaction diagrams, and complex tables or figures. Such information is difficult for general-purpose document parsing systems to directly convert into machin…https://arxiv.org/abs/2607.18839https://arxiv.org/abs/2607.18839Tue, 21 Jul 2026 00:00:00 +0000Shu Wei, Jingjing Wu, Lingshu Zhang, Qunyi Xie, et al. Efficient teamwork typically combines global coordination with parallel execution, a principle not yet fully reflected in unified Vision-Language Model (VLM)-based document parsers. Existing unified parsers process an entire page jointly but generate its output through a single token-by-token autoregressive trajectory,…https://arxiv.org/abs/2607.13639https://arxiv.org/abs/2607.13639Wed, 15 Jul 2026 00:00:00 +0000Shiyin Lu, Yinglun Li, Yu Xia, Yuhui Chen, et al. We introduce OvisOCR2, a 0.8B document parsing model. OvisOCR2 is designed as an end-to-end parser: given a document page image, it generates a Markdown representation in natural reading order, covering text, formulas, tables, and visual regions. We build a data engine that combines filtered real-document annotations w…https://arxiv.org/abs/2607.11562https://arxiv.org/abs/2607.11562Mon, 13 Jul 2026 00:00:00 +0000Yuliang Liu, Zhang Li, Ziyang Zhang, Shuo Zhang, et al. Mainstream visual encoders are pretrained on natural images and cannot be effectively applied to document images without document-oriented adaptation, as dense text and fine-grained character strokes demand character-level visual perception. We present MonkeyOCRv2, a visual-text pretrained model for document AI. First,…https://arxiv.org/abs/2607.07836https://arxiv.org/abs/2607.07836Wed, 08 Jul 2026 00:00:00 +0000Zuming Huang, Jun Huang, Kexuan Ren, Baode Wang, et al. We present Infinity-Parser2, a large multimodal model that couples a controllable data-synthesis pipeline with multi-task reinforcement learning for end-to-end document parsing, addressing the persistent scarcity of faithfully annotated parsing corpora. Our contributions are threefold. First, we build a scalable synthe…https://arxiv.org/abs/2607.02956https://arxiv.org/abs/2607.02956Fri, 03 Jul 2026 00:00:00 +0000Long Xu, Binghong Wu, Tinghao Yu, Hao Feng, et al. · ICML 2026 Multilingual documents encapsulate rich regional cultures, scientific discoveries, and historical records. Parsing this content into structured, machine-readable formats is critical for unlocking global knowledge. However, existing benchmarks predominantly focus on high-resource languages like English and Chinese, crea…https://arxiv.org/abs/2606.24447https://arxiv.org/abs/2606.24447Tue, 23 Jun 2026 00:00:00 +0000Le Xiang, Chenxi Zhai, Shu Wei, Jingjing Wu, et al. Vision-Language Models (VLMs) have revolutionized document parsing by enabling end-to-end mapping from images to structured text, imposing a significant latency bottleneck, particularly for token-dense documents. While Multi-Token Prediction (MTP) has emerged as a promising approach for accelerating inference, its pote…https://arxiv.org/abs/2606.23344https://arxiv.org/abs/2606.23344Mon, 22 Jun 2026 00:00:00 +0000Cheng Cui, Tingquan Gao, Xueqing Wang, Changda Zhou, et al. · ECCV 2026 Accurate document layout analysis remains a critical bottleneck for document parsing systems, due to the intricate coupling among heterogeneous document layout elements, geometric distortions (\eg, paper warping and bending, perspective variations), and reading order within diverse layout structures. Existing approache…https://arxiv.org/abs/2606.12976https://arxiv.org/abs/2606.12976Thu, 11 Jun 2026 00:00:00 +0000Akbar Erkinov, Nurmukhammad Abdurasulov Sharing mathematical content in online forums remains a significant friction point for students and educators: writing raw LATEX is error-prone, standalone optical character recognition tools require platform switching, and current forum software offers no integrated path from a photograph of a formula to a rendered po…https://arxiv.org/abs/2606.11977https://arxiv.org/abs/2606.11977Wed, 10 Jun 2026 00:00:00 +0000LeKai Yu, Hao Liu, Kun Wang, Zhiran Li, et al. In this report, we present our third-place solution for the DataMFM Challenge Track 1: Document Parsing. This track requires models to recover structured Markdown documents from document page images while preserving textual content and document structure. To address the complementary requirements of accurate content re…https://arxiv.org/abs/2606.07401https://arxiv.org/abs/2606.07401Fri, 05 Jun 2026 00:00:00 +0000Ameya Joshi, Joon Kim, Gus Eggert, Joseph Bajor, et al. Document parsing systems are increasingly deployed in high-stakes, regulated workflows such as mortgage underwriting, financial reporting, supply-chain logistics, and clinical records. Yet most public benchmarks evaluate parsers on clean academic layouts or synthetic prose, and report a single OCR or markdown-level sim…https://arxiv.org/abs/2606.07661https://arxiv.org/abs/2606.07661Wed, 03 Jun 2026 00:00:00 +0000Maksim Shandybo, Ivan Bespalov, Daniil Yefimov, Marina Kosheleva, et al. Parsing historical documents with complex, non-standard layouts remains a fundamental bottleneck in large-scale archival digitization. Unlike modern typography, historical newspapers exhibit severe physical degradation and highly irregular page structures that confound even state-of-the-art vision-language models, pres…https://arxiv.org/abs/2606.03264https://arxiv.org/abs/2606.03264Tue, 02 Jun 2026 00:00:00 +0000Zelun Zhang, Hongen Liu, Suyin Liang, Yubo Zhang, et al. We introduce PaddleOCR-VL-1.6, an upgraded compact document parsing model built upon PaddleOCR-VL-1.5. Although PaddleOCR-VL-1.5 establishes a strong 0.9B baseline, its remaining errors concentrate in under-optimized regions where model behavior is unstable, data coverage is sparse, or supervision is unreliable. Rather…https://arxiv.org/abs/2606.01393https://arxiv.org/abs/2606.01393Sun, 31 May 2026 00:00:00 +0000Minglai Yang, Xinyan Velocity Yu, Pengyuan Li, Xinyu Guo, et al. Document parsing and recognition are fundamental capabilities for vision-language models (VLMs) and document processing systems. However, existing Optical Character Recognition (OCR) and document parsing benchmarks are increasingly limited in coverage and difficulty: many focus on common document genres or uniformly sa…https://arxiv.org/abs/2605.24973https://arxiv.org/abs/2605.24973Sun, 24 May 2026 00:00:00 +0000Bangrui Xu, Ziyang Miao, Xuanhe Zhou, Yiming Lin, et al. VLM-based OCR models have become the de facto choice for document parsing, as they can accurately extract page-level elements (e.g., paragraphs within individual pages) together with their bounding boxes and textual content. However, downstream applications such as RAG require coherent document-level information, where…https://arxiv.org/abs/2605.22100https://arxiv.org/abs/2605.22100Thu, 21 May 2026 00:00:00 +0000Bangbang Zhou, Hangdi Xing, Yifan Chen, Jianjun Xu, et al. Document parsing converts visually rich documents into machine-readable structured representations, forming a crucial foundation for information systems. Although many benchmarks have been proposed for document parsing, they remain inadequate for realistic scenarios. Existing benchmarks either focus on specific tasks o…