Pith. sign in

REVIEW 4 major objections 7 minor 76 references

NaviDC-OCR: Navigating Document Parsing Across Digital and Camera-Captured Documents

T0 review · 4 major / 7 minor · reviewed 2026-08-15 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read NaviDC-OCR claims that a single 1.2B vision-language model parses both digital and camera-captured documents at state-of-the-art accuracy, without a separate dewarping module.

desk verdict A credible engineering system with a plausible design story, but the headline SOTA claim is not yet solid: the PureDocBench number rests on an unverified post hoc exclusion, and the 'Degraded track' claim conflicts with the paper's own Table 3. read the letter →

arxiv 2608.12898 v1 pith:YSU7DTWJ submitted 2026-08-13 cs.CV cs.AI

Peng Cai , Zhaofan Zou , Shifa Liu , Yikun Wang , Jiawei Tang , Kaicheng Yang , Meng Tong , Zhongjiang He
show 1 more author
Hao Sun
This is my paper · ORCID
classification cs.CVcs.AI
keywords documentparsingvision-languagemodelscamera-captureddocumentsgeometricdistortionlayoutsegmentationtablestructurerecognitionformulascientificfigure-to-table
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

NaviDC-OCR is a unified document-parsing framework that tries to establish that one model can handle both born-digital and camera-captured documents, including tables, formulas, and scientific figures. The paper's core claim is that a decoupled VLM can beat end-to-end rivals on distorted photos if geometric deformation is learned internally rather than removed by a preprocessing dewarping step. To do this, NaviDC-OCR predicts polygonal region boundaries and a coarse grid of deformation control points, samples boundary points with a curvature-aware simplification rule, and trains tables and formulas by predicting structure before content. On three public benchmarks and one ICDAR competition it reports state-of-the-art overall scores: 96.87 on OmniDocBench v1.6, 88.53 on Wild-OmniDocBench, 78.41 on PureDocBench, and first place in the Sci-ImageMiner figure-to-table task. If these numbers hold, the practical payoff is a single small model that replaces a rectangular-detection-plus-dewarping pipeline with geometry-aware parsing.

What carries the argument

The load-bearing machinery is a set of three training-side mechanisms. Deformation-aware learning turns synthetic warps into two supervision signals: region-level polygon boundaries (replacing rectangular boxes) and a downsampled $32\times 32$ grid of control points that the VLM must predict, with the warps generated through forward mappings from Doc3D and ForCenNet. Curvature-Guided Douglas–Peucker Sampling (CGDP) assigns each candidate boundary point an importance $S_i = d_i(1 + \lambda\hat{\kappa}_i)$, mixing global simplification distance with local curvature so creases and sharp corners keep enough sample points. Content-structure decoupled learning makes the model emit a structural skeleton first—OTSL table topology without cell contents, or LaTeX syntax tokens—and then fill in content, which the paper argues lowers structural uncertainty. Around these sit a data engine (heterogeneous-model consensus voting plus a self-judgement VLM that compares rendered predictions to the original image) and a four-stage training schedule ending in GRPO reinforcement learning with normalized edit distance, TEDS, and CDM rewards.

What would settle it

Train NaviDC-OCR twice, once with and once without the Stage 2 synthetic-deformation samples, keeping everything else identical, and compare Wild-OmniDocBench scores on creased and folded pages; if the gap is within noise, the deformation transfer claim fails. As a sharper check, hold out real photos with warp types absent from the Doc3D/ForCenNet synthesis, such as page curl from bound books or water-ripple distortion, and see whether accuracy drops.

Watch

Extended reading notes

Core claim

The central claim is that document geometry can be absorbed into a vision-language model as an intermediate prediction target, so that camera-captured documents no longer need explicit rectification. The authors report that applying a dewarping model to Wild-OmniDocBench substantially improves a two-stage parser, and that structure-heavy tasks such as tables and formulas show high prediction entropy; NaviDC-OCR is their answer to both. Layout analysis is reformulated as polygon boundary prediction with an adaptive number of points, supplemented by a coarse global control-point field that teaches the VLM the page's warp, and structured tasks are taught in two passes: first the OTSL table skeleton or LaTeX syntax, then the cell contents. The trained 1.2B model reports 96.87 overall on OmniDocBench v1.6, 88.53 on Wild-OmniDocBench v1.5, 78.41 on PureDocBench, and the best TEDS score in the ICDAR 2026 Sci-ImageMiner Challenge. Taken at face value, the discovery is that deformation-aware, structure-decoupled training makes a decoupled VLM competitive across clean, degraded, and physically captured documents.

Load-bearing premise

The main assumption is that the computer-generated wrinkles, curves, and folds used in training stand in for the distortions in real camera-captured test photos; if that synthetic-to-real transfer fails, the Wild-OmniDocBench advantage and the deformation-aware design are not established.

Editorial extensions

If this is right

  • If the reported scores replicate, decoupled VLM parsers can drop the dedicated dewarping preprocessing module; geometric robustness comes from training on synthetic warps.
  • Polygon-based layout prediction with adaptive sampling becomes a viable replacement for axis-aligned boxes on curved, folded, or creased pages, improving crops for downstream content recognition.
  • The structure-then-content recipe generalizes beyond tables and formulas to any task with a strong grammar, such as code blocks and scientific figure-to-table conversion, the paper suggests.
  • The data pipeline's consensus voting plus rendered-image self-judgement offers a way to build large document-parsing training sets with fewer human annotations.
  • Task-specific verifiable rewards in GRPO provide a template for aligning parsing models to text, table, and formula metrics simultaneously.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • An implication the authors leave implicit is that the deformation-aware recipe (adaptive polygons plus a coarse warp field) could transfer to other distorted-perception tasks such as camera-captured forms, receipts, or whiteboards, where rectangular detectors fail the same way.
  • A testable extension is to isolate CGDP: swapping the curvature modulation $\lambda\hat{\kappa}_i$ off to plain Douglas–Peucker should specifically hurt crease and corner regions in layout segmentation, which would confirm the sampling rule as the active ingredient.
  • The paper's own admission of repetitive generation on several PureDocBench pages points to a follow-up: page-level structure conditioning or constrained decoding for long structured documents, rather than deleting invalid outputs.
  • Because the model already predicts a coarse $32\times 32$ control-point grid, the same supervision could be extended to recover a backward mapping and produce rectified document images as a byproduct, something NaviDC-OCR does not currently output.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 7 minor

Summary. The paper presents NaviDC-OCR, a 1.2B-parameter decoupled vision-language model for unified document parsing of digital and camera-captured documents. The method combines three claimed contributions: deformation-aware learning that replaces rectangular layout detection with polygon segmentation and adds point-level geometric supervision; a Curvature-Guided Douglas–Peucker (CGDP) sampling mechanism for adaptive polygon point selection; and a content-structure decoupled learning strategy that explicitly models formula grammars and table structures. Training is organized in four progressive stages, with multi-node consensus voting and a self-judgement VLM for pseudo-label filtering. The authors report state-of-the-art overall scores of 96.87 on OmniDocBench v1.6, 88.53 on Wild-OmniDocBench, 78.41 on PureDocBench, and first place in the ICDAR 2026 Sci-ImageMiner Challenge. The evaluation uses public benchmarks with a fixed OmniDocBench-style protocol.

Significance. If the reported results hold, NaviDC-OCR would be a meaningful advance in practical document parsing, particularly for camera-captured documents where decoupled pipelines traditionally struggle. The paper offers a detailed data-engineering pipeline, a unified prompt interface for eight parsing tasks, and competitive results against much larger models (e.g., 235B general VLMs and 30B Ovis2.6). The authors explicitly disclose the removal of six invalid PureDocBench prediction files in Appendix D, which is a degree of transparency that should be credited. However, the significance is conditional: the central claim of state-of-the-art performance rests on benchmark tables without ablations that isolate the three proposed contributions, without error bars or significance tests, and with an unresolved issue concerning the handling of missing prediction files in the PureDocBench scoring. The synthetic-to-real transfer assumption behind the deformation-aware learning is also not directly validated.

major comments (4)
  1. [§4, Tables 1–3] The paper attributes the performance gains to deformation-aware learning, CGDP sampling, and content-structure decoupled learning, but no ablation experiment isolates any of these components. For example, there is no comparison of Stage 2 training with and without the point-level deformation supervision, nor a comparison of polygon boundaries produced by CGDP versus uniform sampling, nor a removal of the content-structure decoupled objective in Stage 3. Without such ablations, the headline improvements over OvisOCR2, PaddleOCR-VL-1.6, and MinerU2.5-Pro cannot be attributed to the proposed mechanisms rather than to the larger or better-curated training data, the multi-model pseudo-labeling, or the four-stage training schedule. I request a dedicated ablation table, at minimum on OmniDocBench v1.6 and Wild-OmniDocBench, that turns off each contribution individually.
  2. [Appendix D, Table 5] The PureDocBench overall score of 78.41 is load-bearing for the abstract's and conclusion's SOTA claims, yet it depends on the post hoc removal of six invalid Markdown prediction files. The paper asserts that 'missing predictions are assigned a score of zero under the evaluation protocol,' but this behavior is not verified against the evaluator source code, and the appendix does not demonstrate what the official OmniDocBench quick_match evaluator does with an absent prediction file. If the evaluator silently drops unmatched pages from the denominator rather than scoring them as zero, or if the malformed files would have received partial credit, the reported 78.41 would be inflated. The authors must either (a) provide source-level verification of the zero-missing behavior, or (b) re-score with the six files replaced by empty predictions and report both the original and corrected numbers. The claim in the current form is not verifiable from the manuscript.
  3. [§5.2, Table 3] The sentence 'NaviDC-OCR achieves state-of-the-art performance on the Degraded track in Table 3, outperforming existing end-to-end document parsing methods' is contradicted by the 'Digital Degraded' column, where OvisOCR2 scores 77.77 overall versus NaviDC-OCR's 77.47. If the intended claim is limited to the 'Real Degraded' track, the text should say so explicitly; otherwise, the statement overstates the results. Similar caution is needed for the PureDocBench-Clean discussion in §5.1, where the claim of surpassing OvisOCR2 is supported (86.90 vs. 82.14) and should be kept distinct from the degraded-track claim.
  4. [§3.2.1, §4.2, Table 2] The central innovation, deformation-aware learning, relies on a transfer assumption that is not tested. Synthetic warps are generated by applying Doc3D/ForCenNet forward mappings to undistorted pages and training the model to predict the warped control points and region boundaries. There is no quantitative analysis of the distribution gap between these synthetic deformations and the real capture conditions in Wild-OmniDocBench, and no ablation that removes the deformation-aware supervision while keeping the additional synthetic data volume fixed. Consequently, the improvement on Wild-OmniDocBench (Table 2) could in principle come from the sheer increase in training data rather than from the geometric awareness the paper claims. I would like to see a direct comparison on Wild-OmniDocBench between (i) the full Stage 2 pipeline, (ii) the same data volume without the control-point and polygon supervision, and (iii) a version trained only on digital documents, to establish that the deformation-specific losses are what drive the gain.
minor comments (7)
  1. [Table 2] In the OvisOCR2 row, the numeric entries '0.12990.3785.13' appear to be missing separators and are inconsistent with the column structure; this looks like a formatting error that should be corrected.
  2. [Table 1] The HunyuanOCR-1.5 row lists the exact same Formula CDM, Table TEDS, Table TEDS-S, and Read Order Edit values as the PaddleOCR-VL-1.6 row; please verify whether this is a copy-and-paste error or genuinely identical measurements.
  3. [§4] The paper states that 'we release the complete implementation and model configurations,' but no repository URL, model checkpoint link, or contact for code release is given anywhere in the text; please add the release location.
  4. [§3.2.1 and §4.2] The notation for the number of deformation control points is inconsistent: §3.2.1 introduces 'M×M control points P' while §4.2 uses 'N^2 = 1,024 control points'; unify the notation and clarify whether M and N refer to the same quantity.
  5. [Abstract and §5.2] The abstract refers to 'Wild-OmniDocBench' without a version, while §5.2 and Table 2 specify 'v1.5'; harmonize the version information across the paper.
  6. [§3.2.2] The approximation 'd ≈ κL^2/8' for the Douglas–Peucker distance is stated without derivation or a citation; please provide a reference or a brief derivation, and define κ and L precisely.
  7. [Figure 2] In the upper-left panel of Figure 2, the label 'NaviO R' appears garbled and should be corrected to a readable model name.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity found; all headline results are external benchmark measurements.

full rationale

The paper's central claim is an empirical comparison on independently maintained public benchmarks: OmniDocBench v1.6, Wild-OmniDocBench, PureDocBench, and the ICDAR 2026 Sci-ImageMiner Challenge. The reported scores (96.87, 88.53, 78.41, and first place) are measurements against external ground truth, not quantities derived from the model's own training inputs by construction. No equation in the paper defines a predicted metric in terms of a fitted parameter, and no benchmark score is used as a training target. The deformation-aware training uses synthetic deformations generated with the authors' own ForCenNet forward mappings (Section 3.2.1), which introduces a self-citation and a possible distribution-transfer risk, but this does not make the Wild-OmniDocBench outcome true by construction; the test distribution and scorer are external. The reinforcement learning stage optimizes NED, TEDS, and CDM rewards (Equation 5), which match the evaluation metrics, but optimizing an external metric is not circular unless the test set itself is used for training or the metric is fitted to the model, neither of which is claimed. Appendix D discloses that six invalid PureDocBench prediction files were removed before scoring, with the assertion that missing predictions receive zero; this is a scoring-robustness concern rather than a circularity, because removing a file that would score zero does not by itself force the reported number. Overall, the central claims are supported by independent benchmark comparisons, and no load-bearing step reduces to its own inputs by definition or by self-citation chain.

Assumptions & free parameters 4 free parameters · 5 assumptions · 0 invented entities

No new physical entities are introduced. The Self-Judgement VLM and CGDP sampling are trained procedures, not postulated entities. The paper's load-bearing supports are data-generation and training assumptions rather than new objects.

free parameters (4)
  • lambda in CGDP (curvature modulation weight)
    Equation (3), S_i = d_i (1 + lambda * kappa_hat_i), controls how much local curvature shifts Douglas-Peucker sampling priority; no value or ablation is reported.
  • tau for MCV consensus threshold
    Section 3.1: a prediction is accepted only if max_i C_i >= tau; the threshold controls training data volume versus precision and is not specified or ablated.
  • tau for CGDP selection threshold
    Section 3.2.2: 'When max_i S_i > tau, the corresponding point is selected'; this threshold determines polygon density and is not specified.
  • N^2 = 1,024 point-level deformation control points = 1024
    Stage 2 (Section 4.2) downsamples deformation fields to 32x32 control points instead of 82,944 used by ForCenNet; this design choice defines the deformation representation capacity.
assumptions (5)
  • domain assumption Consensus Hypothesis: for heterogeneous models, predictions supported by multiple models are more likely to be reliable than individual predictions.
    Section 3.1 grounds pseudo-label selection; if all voting models share systematic biases, the consensus labels inherit them and the training data quality claim collapses.
  • domain assumption Synthetic deformations produced by Doc3D/ForCenNet forward mappings approximate real camera-captured document distortions.
    Section 3.2.1 and Stage 2 use these mappings to create camera-distorted training samples; the claimed transfer to Wild-OmniDocBench depends on distributional realism.
  • domain assumption For smooth curves the Douglas-Peucker distance approximately satisfies d approx kappa L^2 / 8.
    Section 3.2.2 uses this relation to justify combining DP distance with curvature; the paper gives no derivation or verification of the constant.
  • domain assumption A fine-tuned 7B self-judgement VLM can reliably filter and correct pseudo-labels via image-to-image comparison.
    Section 3.3 shows a 235B general VLM has recall below 40% and then assumes the fine-tuned Qwen2.5-VL-7B judger is sufficient; no precision or recall for J_theta is reported.
  • domain assumption OmniDocBench-style metrics and the overall score formula adequately measure document parsing quality.
    Appendix D defines Overall as the average of text, formula, and table scores; the SOTA claim is relative to this specific protocol and may not generalize to other evaluation standards.

how reviews work

0 comments
Cite this review

Pith. "Pith review of NaviDC-OCR: Navigating Document Parsing Across Digital and Camera-Captured Documents." pith.science (2026). https://pith.science/paper/YSU7DTWJ

@misc{pith2026260812898,
  author       = {Pith},
  title        = {Pith review of: NaviDC-OCR: Navigating Document Parsing Across Digital and Camera-Captured Documents},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/YSU7DTWJ}},
  note         = {Machine review of arXiv:2608.12898}
}
read the original abstract

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 accurate layout analysis, where geometric distortions in camera-captured documents can introduce cascading errors. Second, although end-to-end VLM-based methods alleviate the dependence on explicit layout detection, they often suffer from redundant generation, hallucinations, and insufficient structural reasoning in high-resolution scenarios. To address these challenges, we propose NaviDC-OCR, a unified framework for document parsing. NaviDC-OCR introduces deformation-aware learning to incorporate geometric perception into VLMs and proposes an adaptive sampling mechanism for complex layout representation. Furthermore, a content-structure decoupled learning strategy is developed to explicitly model formula grammars and table structures, enabling more effective structured representation learning. Extensive experiments demonstrate that NaviDC-OCR achieves state-of-the-art performance across diverse document parsing benchmarks. It obtains overall scores of 96.87, 88.53 and 78.41 on OmniDocBench v1.6, Wild-OmniDocBench, and PureDocBench, respectively, and ranks first in the ICDAR 2026 Sci-ImageMiner Challenge. These results validate the effectiveness and generalization capability of NaviDC-OCR in complex document parsing scenarios.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

76 extracted references · 35 canonical work pages

  1. [3]

    Qwen2.5-vl technical report, 2025

    Shuai Bai, Keqin Chen, Xuejing Liu, Jialin Wang, Wenbin Ge, Sibo Song, Kai Dang, Peng Wang, Shijie Wang, Jun Tang, Humen Zhong, Yuanzhi Zhu, Mingkun Yang, Zhaohai Li, Jianqiang Wan, Pengfei Wang, Wei Ding, Zheren Fu, Yiheng Xu, Jiabo Ye, Xi Zhang, Tianbao Xie, Zesen Cheng, Hang Zhang, Zhibo Yang, Haiyang Xu, and Junyang Lin. Qwen2.5-vl technical report, 2...

    arXiv 2025

  2. [4]

    Forcennet: Foreground-centric network for document image rectification

    Peng Cai, Qiang Li, Kaicheng Yang, Dong Guo, Jia Li, Nan Zhou, Xiang An, Ninghua Yang, and Jiankang Deng. Forcennet: Foreground-centric network for document image rectification. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 15137--15146, 2025

    work page 2025

  3. [6]

    Sam4mllm: Enhance multi-modal large language model for referring expression segmentation

    Yi-Chia Chen, Wei-Hua Li, Cheng Sun, Yu-Chiang Frank Wang, and Chu-Song Chen. Sam4mllm: Enhance multi-modal large language model for referring expression segmentation. In European Conference on Computer Vision, pages 323--340. Springer, 2024

    work page 2024

  4. [9]

    Dewarpnet: Single-image document unwarping with stacked 3d and 2d regression networks

    Sagnik Das, Ke Ma, Zhixin Shu, Dimitris Samaras, and Roy Shilkrot. Dewarpnet: Single-image document unwarping with stacked 3d and 2d regression networks. In Proceedings of the IEEE/CVF international conference on computer vision, pages 131--140, 2019

    work page 2019

  5. [12]

    Glm-ocr technical report, 2026

    Shuaiqi Duan, Yadong Xue, Weihan Wang, Zhe Su, Huan Liu, Sheng Yang, Guobing Gan, Guo Wang, Zihan Wang, Shengdong Yan, Dexin Jin, Yuxuan Zhang, Guohong Wen, Yanfeng Wang, Yutao Zhang, Xiaohan Zhang, Wenyi Hong, Yukuo Cen, Da Yin, Bin Chen, Wenmeng Yu, Xiaotao Gu, and Jie Tang. Glm-ocr technical report, 2026. URL https://arxiv.org/abs/2603.10910

    arXiv 2026

  6. [13]

    Geometric representation learning for document image rectification

    Hao Feng, Wengang Zhou, Jiajun Deng, Yuechen Wang, and Houqiang Li. Geometric representation learning for document image rectification. In ECCV, pages 475--492. Springer, 2022

    work page 2022

  7. [14]

    Dolphin: Document image parsing via heterogeneous anchor prompting, 2025

    Hao Feng, Shu Wei, Xiang Fei, Wei Shi, Yingdong Han, Lei Liao, Jinghui Lu, Binghong Wu, Qi Liu, Chunhui Lin, Jingqun Tang, Hao Liu, and Can Huang. Dolphin: Document image parsing via heterogeneous anchor prompting, 2025. URL https://arxiv.org/abs/2505.14059

    arXiv 2025

  8. [16]

    Speeding up the douglas-peucker line-simplification algorithm

    John Edward Hershberger and Jack Snoeyink. Speeding up the douglas-peucker line-simplification algorithm. 1992

    work page 1992

Show all 76 references
  1. [17]

    Revisiting document image dewarping by grid regularization

    Xiangwei Jiang, Rujiao Long, Nan Xue, Zhibo Yang, Cong Yao, and Gui-Song Xia. Revisiting document image dewarping by grid regularization. In CVPR, pages 4543--4552, 2022

    2022

  2. [23]

    Ovis: Structural embedding alignment for multimodal large language model, 2024

    Shiyin Lu, Yang Li, Qing-Guo Chen, Zhao Xu, Weihua Luo, Kaifu Zhang, and Han-Jia Ye. Ovis: Structural embedding alignment for multimodal large language model, 2024. URL https://arxiv.org/abs/2405.20797

    2024 arXiv

  3. [24]

    Ovis2.5 technical report

    Shiyin Lu, Yang Li, Yu Xia, Yuwei Hu, Shanshan Zhao, Yanqing Ma, Zhichao Wei, Yinglun Li, Lunhao Duan, Jianshan Zhao, Yuxuan Han, Haijun Li, Wanying Chen, Junke Tang, Chengkun Hou, Zhixing Du, Tianli Zhou, Wenjie Zhang, Huping Ding, Jiahe Li, Wen Li, Gui Hu, Yiliang Gu, Siran ...

    2025 arXiv

  4. [26]

    Optimized table tokenization for table structure recognition

    Maksym Lysak, Ahmed Nassar, Nikolaos Livathinos, Christoph Auer, and Peter Staar. Optimized table tokenization for table structure recognition. In International Conference on Document Analysis and Recognition, pages 37--50. Springer, 2023

    2023

  5. [28]

    Omnidocbench: Benchmarking diverse pdf document parsing with comprehensive annotations, 2024

    Linke Ouyang, Yuan Qu, Hongbin Zhou, Jiawei Zhu, Rui Zhang, Qunshu Lin, Bin Wang, Zhiyuan Zhao, Man Jiang, Xiaomeng Zhao, Jin Shi, Fan Wu, Pei Chu, Minghao Liu, Zhenxiang Li, Chao Xu, Bo Zhang, Botian Shi, Zhongying Tu, and Conghui He. Omnidocbench: Benchmarking diverse pdf do...

    2024 arXiv

  6. [29]

    Perceptiongpt: Effectively fusing visual perception into llm

    Renjie Pi, Lewei Yao, Jiahui Gao, Jipeng Zhang, and Tong Zhang. Perceptiongpt: Effectively fusing visual perception into llm. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 27124--27133, 2024

    2024

  7. [30]

    Jack of all tasks master of many: Designing general-purpose coarse-to-fine vision-language model

    Shraman Pramanick, Guangxing Han, Rui Hou, Sayan Nag, Ser-Nam Lim, Nicolas Ballas, Qifan Wang, Rama Chellappa, and Amjad Almahairi. Jack of all tasks master of many: Designing general-purpose coarse-to-fine vision-language model. In Proceedings of the IEEE/CVF Conference on Co...

    2024

  8. [32]

    Ufo: A unified approach to fine-grained visual perception via open-ended language interface

    Hao Tang, Chen-Wei Xie, Haiyang Wang, Xiaoyi Bao, Tingyu Weng, Pandeng Li, Yun Zheng, and Liwei Wang. Ufo: A unified approach to fine-grained visual perception via open-ended language interface. Advances in Neural Information Processing Systems, 38: 0 83761--83791, 2026

    2026

  9. [33]

    Hunyuanocr technical report, 2025

    Hunyuan Vision Team, Pengyuan Lyu, Xingyu Wan, Gengluo Li, Shangpin Peng, Weinong Wang, Liang Wu, Huawen Shen, Yu Zhou, Canhui Tang, Qi Yang, Qiming Peng, Bin Luo, Hower Yang, Xinsong Zhang, Jinnian Zhang, Houwen Peng, Hongming Yang, Senhao Xie, Longsha Zhou, Ge Pei, Binghong ...

    2025

  10. [35]

    Image over text: Transforming formula recognition evaluation with character detection matching

    Bin Wang, Fan Wu, Linke Ouyang, Zhuangcheng Gu, Rui Zhang, Renqiu Xia, Botian Shi, Bo Zhang, and Conghui He. Image over text: Transforming formula recognition evaluation with character detection matching. In 2025 IEEE/CVF Conference on Computer Vision and Pattern Recognition (...

    2025

  11. [38]

    Internvl3.5: Advancing open-source multimodal models in versatility, reasoning, and efficiency, 2025 c

    Weiyun Wang, Zhangwei Gao, Lixin Gu, Hengjun Pu, Long Cui, Xingguang Wei, Zhaoyang Liu, Linglin Jing, Shenglong Ye, Jie Shao, Zhaokai Wang, Zhe Chen, Hongjie Zhang, Ganlin Yang, Haomin Wang, Qi Wei, Jinhui Yin, Wenhao Li, Erfei Cui, Guanzhou Chen, Zichen Ding, Changyao Tian, Z...

    2025 arXiv

  12. [40]

    Firered-ocr technical report, 2026

    Hao Wu, Haoran Lou, Xinyue Li, Zuodong Zhong, Zhaojun Sun, Phellon Chen, Xuanhe Zhou, Kai Zuo, Yibo Chen, Xu Tang, Yao Hu, Boxiang Zhou, Jian Wu, Yongji Wu, Wenxin Yu, Yingmiao Liu, Yuhao Huang, Manjie Xu, Gang Liu, Yidong Ma, Zhichao Sun, and Changhao Qiao. Firered-ocr techni...

    2026

  13. [41]

    Document dewarping with control points

    Guo-Wang Xie, Fei Yin, Xu-Yao Zhang, and Cheng-Lin Liu. Document dewarping with control points. In ICDAR, pages 466--480. Springer, 2021

    2021

  14. [44]

    Youtu-parsing: Perception, structuring and recognition via high-parallelism decoding, 2026

    Kun Yin, Yunfei Wu, Bing Liu, Zhongpeng Cai, Xiaotian Li, Huang Chen, Xin Li, Haoyu Cao, Yinsong Liu, Deqiang Jiang, Xing Sun, Yunsheng Wu, Qianyu Li, Antai Guo, Yanzhen Liao, Yanqiu Qu, Haodong Lin, Chengxu He, and Shuangyin Liu. Youtu-parsing: Perception, structuring and rec...

    2026 arXiv

  15. [48]

    Image-based table recognition: data, model, and evaluation

    Xu Zhong, Elaheh ShafieiBavani, and Antonio Jimeno Yepes. Image-based table recognition: data, model, and evaluation. In European conference on computer vision, pages 564--580. Springer, 2020

    2020

  16. [50]

    Reading or reasoning? format decoupled reinforcement learning for document ocr

    Yufeng Zhong, Lei Chen, Zhixiong Zeng, Xuanle Zhao, Deyang Jiang, Liming Zheng, Jing Huang, Haibo Qiu, Peng Shi, Siqi Yang, et al. Reading or reasoning? format decoupled reinforcement learning for document ocr. In Proceedings of the IEEE/CVF Conference on Computer Vision and P...

    2026

  17. [51]

    Ocrverse: Towards holistic ocr in end-to-end vision-language models, 2026 b

    Yufeng Zhong, Lei Chen, Xuanle Zhao, Wenkang Han, Liming Zheng, Jing Huang, Deyang Jiang, Yilin Cao, Lin Ma, and Zhixiong Zeng. Ocrverse: Towards holistic ocr in end-to-end vision-language models, 2026 b . URL https://arxiv.org/abs/2601.21639

    2026

  18. [52]

    arXiv preprint arXiv:2510.12323 , year=

    Rag-anything: All-in-one rag framework , author=. arXiv preprint arXiv:2510.12323 , year=

  19. [53]

    arXiv preprint arXiv:2512.03413 , year=

    BookRAG: A Hierarchical Structure-aware Index-based Approach for Retrieval-Augmented Generation on Complex Documents , author=. arXiv preprint arXiv:2512.03413 , year=

  20. [54]

    5-Pro: Pushing the limits of data-centric document parsing at scale , author=

    MinerU2. 5-Pro: Pushing the limits of data-centric document parsing at scale , author=. arXiv preprint arXiv:2604.04771 , year=

    arXiv

  21. [55]

    arXiv preprint arXiv:2603.23885 , year=

    Towards real-world document parsing via realistic scene synthesis and document-aware training , author=. arXiv preprint arXiv:2603.23885 , year=

    arXiv

  22. [56]

    Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

    Forcennet: Foreground-centric network for document image rectification , author=. Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

  23. [57]

    arXiv preprint arXiv:2601.08834 , year=

    Reading or Reasoning? Format Decoupled Reinforcement Learning for Document OCR , author=. arXiv preprint arXiv:2601.08834 , year=

  24. [58]

    Proceedings of the IEEE/CVF international conference on computer vision , pages=

    Dewarpnet: Single-image document unwarping with stacked 3d and 2d regression networks , author=. Proceedings of the IEEE/CVF international conference on computer vision , pages=

  25. [59]

    arXiv preprint arXiv:2601.21957 , year=

    PaddleOCR-VL-1.5: Towards a Multi-Task 0.9 B VLM for Robust In-the-Wild Document Parsing , author=. arXiv preprint arXiv:2601.21957 , year=

    arXiv

  26. [60]

    2025 , eprint=

    Qwen2.5-VL Technical Report , author=. 2025 , eprint=

    2025

  27. [61]

    arXiv preprint arXiv:2505.09388 , year=

    Qwen3 technical report , author=. arXiv preprint arXiv:2505.09388 , year=

    arXiv

  28. [62]

    arXiv preprint arXiv:2607.13639 , year=

    OvisOCR2 Technical Report , author=. arXiv preprint arXiv:2607.13639 , year=

    arXiv

  29. [63]

    arXiv preprint arXiv:2510.18234 , year=

    Deepseek-ocr: Contexts optical compression , author=. arXiv preprint arXiv:2510.18234 , year=

    arXiv

  30. [64]

    2025 , eprint=

    HunyuanOCR Technical Report , author=. 2025 , eprint=

    2025

  31. [65]

    arXiv preprint arXiv:2603.09677 , year=

    Logics-parsing-omni technical report , author=. arXiv preprint arXiv:2603.09677 , year=

    arXiv

  32. [66]

    2025 , eprint=

    Dolphin: Document Image Parsing via Heterogeneous Anchor Prompting , author=. 2025 , eprint=

    2025

  33. [67]

    5 Technical Report: Unlocking Robust Document Parsing for Complex Patterns , author=

    MonkeyOCR v1. 5 Technical Report: Unlocking Robust Document Parsing for Complex Patterns , author=. arXiv preprint arXiv:2511.10390 , year=

  34. [68]

    2026 , eprint=

    GLM-OCR Technical Report , author=. 2026 , eprint=

    2026

  35. [69]

    2026 , eprint=

    Youtu-Parsing: Perception, Structuring and Recognition via High-Parallelism Decoding , author=. 2026 , eprint=

    2026

  36. [70]

    arXiv preprint arXiv:2606.03264 , year=

    PaddleOCR-VL-1.6: Expanding the Frontier of Document Parsing with Under-Optimized Region Refinement and Progressive Post-Training , author=. arXiv preprint arXiv:2606.03264 , year=

    arXiv

  37. [71]

    ICDAR , pages=

    Document dewarping with control points , author=. ICDAR , pages=. 2021 , organization=

    2021

  38. [72]

    ECCV , pages=

    Geometric representation learning for document image rectification , author=. ECCV , pages=. 2022 , organization=

    2022

  39. [73]

    CVPR , pages=

    Revisiting document image dewarping by grid regularization , author=. CVPR , pages=

  40. [74]

    European Conference on Computer Vision , pages=

    Sam4mllm: Enhance multi-modal large language model for referring expression segmentation , author=. European Conference on Computer Vision , pages=. 2024 , organization=

    2024

  41. [75]

    arXiv preprint arXiv:2511.16719 , year=

    Sam 3: Segment anything with concepts , author=. arXiv preprint arXiv:2511.16719 , year=

    arXiv

  42. [76]

    Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

    Perceptiongpt: Effectively fusing visual perception into llm , author=. Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

  43. [77]

    arXiv preprint arXiv:2605.07141 , year=

    Qwen3-VL-Seg: Unlocking Open-World Referring Segmentation with Vision-Language Grounding , author=. arXiv preprint arXiv:2605.07141 , year=

    arXiv

  44. [78]

    Advances in Neural Information Processing Systems , volume=

    Ufo: A unified approach to fine-grained visual perception via open-ended language interface , author=. Advances in Neural Information Processing Systems , volume=

  45. [79]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    Jack of all tasks master of many: Designing general-purpose coarse-to-fine vision-language model , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  46. [80]

    arXiv preprint arXiv:2605.07492 , year=

    How Far Is Document Parsing from Solved? PureDocBench: A Source-TraceableBenchmark across Clean, Degraded, and Real-World Settings , author=. arXiv preprint arXiv:2605.07492 , year=

    arXiv

  47. [81]

    International Conference on Document Analysis and Recognition , pages=

    Optimized table tokenization for table structure recognition , author=. International Conference on Document Analysis and Recognition , pages=. 2023 , organization=

    2023

  48. [82]

    1992 , publisher=

    Speeding up the Douglas-Peucker line-simplification algorithm , author=. 1992 , publisher=

    1992

  49. [83]

    arXiv preprint arXiv:2402.03300 , year=

    Deepseekmath: Pushing the limits of mathematical reasoning in open language models , author=. arXiv preprint arXiv:2402.03300 , year=

    arXiv

  50. [84]

    arXiv preprint arXiv:2510.14528 , year=

    Paddleocr-vl: Boosting multilingual document parsing via a 0.9 b ultra-compact vision-language model , author=. arXiv preprint arXiv:2510.14528 , year=

  51. [85]

    2026 , eprint=

    FireRed-OCR Technical Report , author=. 2026 , eprint=

    2026

  52. [86]

    arXiv preprint arXiv:2509.22186 , year=

    MinerU2.5: A Decoupled Vision-Language Model for Efficient High-Resolution Document Parsing , author=. arXiv preprint arXiv:2509.22186 , year=

    arXiv

  53. [87]

    arXiv preprint arXiv:2512.21095 , year=

    UniRec-0.1B: Unified Text and Formula Recognition with 0.1B Parameters , author=. arXiv preprint arXiv:2512.21095 , year=

    arXiv

  54. [88]

    ocr: Multilingual document layout parsing in a single vision-language model , author=

    dots. ocr: Multilingual document layout parsing in a single vision-language model , author=. arXiv preprint arXiv:2512.02498 , year=

  55. [89]

    2026 , eprint=

    OCRVerse: Towards Holistic OCR in End-to-End Vision-Language Models , author=. 2026 , eprint=

    2026

  56. [90]

    arXiv preprint arXiv:2506.05218 , year=

    MonkeyOCR: Document Parsing with a Structure-Recognition-Relation Triplet Paradigm , author=. arXiv preprint arXiv:2506.05218 , year=

  57. [91]

    2024 , eprint=

    Ovis: Structural Embedding Alignment for Multimodal Large Language Model , author=. 2024 , eprint=

    2024

  58. [92]

    2025 , journal=

    Ovis2.5 Technical Report , author=. 2025 , journal=

    2025

  59. [93]

    2025 , eprint=

    InternVL3.5: Advancing Open-Source Multimodal Models in Versatility, Reasoning, and Efficiency , author=. 2025 , eprint=

    2025

  60. [94]

    arXiv preprint arXiv:2607.04884 , year=

    HunyuanOCR-1.5: Making Lightweight OCR VLMs Faster and Better , author=. arXiv preprint arXiv:2607.04884 , year=

    arXiv

  61. [95]

    arXiv preprint arXiv:2603.13032 , year=

    Multimodal OCR: Parse anything from documents , author=. arXiv preprint arXiv:2603.13032 , year=

  62. [96]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    Reading or reasoning? format decoupled reinforcement learning for document ocr , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  63. [97]

    arXiv preprint arXiv:2603.13398 , year=

    Qianfan-OCR: A unified end-to-end model for document intelligence , author=. arXiv preprint arXiv:2603.13398 , year=

  64. [98]

    5-omni technical report , author=

    Qwen3. 5-omni technical report , author=. arXiv preprint arXiv:2604.15804 , year=

    arXiv

  65. [99]

    2025 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , pages=

    Image over text: Transforming formula recognition evaluation with character detection matching , author=. 2025 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , pages=. 2025 , organization=

    2025

  66. [100]

    European conference on computer vision , pages=

    Image-based table recognition: data, model, and evaluation , author=. European conference on computer vision , pages=. 2020 , organization=

    2020

  67. [101]

    2024 , eprint=

    OmniDocBench: Benchmarking Diverse PDF Document Parsing with Comprehensive Annotations , author=. 2024 , eprint=

    2024

  68. [102]

    arXiv preprint arXiv:2607.26848 , year=

    ICDAR 2026 Competition on Information Extraction from Atomic Layer Deposition/Etching (ALD/E) Scientific Figures , author=. arXiv preprint arXiv:2607.26848 , year=

    2026 arXiv

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.