Saved in:
| Main Authors: | , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2510.15349 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866911219568345088 |
|---|---|
| author | Wang, Baode Wu, Biao Li, Weizhen Fang, Meng Huang, Zuming Huang, Jun Wang, Haozhe Liang, Yanjie Chen, Ling Chu, Wei Qi, Yuan |
| author_facet | Wang, Baode Wu, Biao Li, Weizhen Fang, Meng Huang, Zuming Huang, Jun Wang, Haozhe Liang, Yanjie Chen, Ling Chu, Wei Qi, Yuan |
| contents | Document parsing from scanned images into structured formats remains a significant challenge due to its complexly intertwined elements such as text paragraphs, figures, formulas, and tables. Existing supervised fine-tuning methods often struggle to generalize across diverse document types, leading to poor performance, particularly on out-of-distribution data. This issue is further exacerbated by the limited availability of high-quality training data for layout-aware parsing tasks. To address these challenges, we introduce LayoutRL, a reinforcement learning framework that optimizes layout understanding through composite rewards integrating normalized edit distance, paragraph count accuracy, and reading order preservation. To support this training, we construct the Infinity-Doc-400K dataset, which we use to train Infinity-Parser, a vision-language model demonstrating robust generalization across various domains. Extensive evaluations on benchmarks including OmniDocBench, olmOCR-Bench, PubTabNet, and FinTabNet show that Infinity-Parser consistently achieves state-of-the-art performance across a broad range of document types, languages, and structural complexities, substantially outperforming both specialized document parsing systems and general-purpose vision-language models. We will release our code, dataset, and model to facilitate reproducible research in document parsing. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2510_15349 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Infinity Parser: Layout Aware Reinforcement Learning for Scanned Document Parsing Wang, Baode Wu, Biao Li, Weizhen Fang, Meng Huang, Zuming Huang, Jun Wang, Haozhe Liang, Yanjie Chen, Ling Chu, Wei Qi, Yuan Computation and Language F.2.2; I.2.7 Document parsing from scanned images into structured formats remains a significant challenge due to its complexly intertwined elements such as text paragraphs, figures, formulas, and tables. Existing supervised fine-tuning methods often struggle to generalize across diverse document types, leading to poor performance, particularly on out-of-distribution data. This issue is further exacerbated by the limited availability of high-quality training data for layout-aware parsing tasks. To address these challenges, we introduce LayoutRL, a reinforcement learning framework that optimizes layout understanding through composite rewards integrating normalized edit distance, paragraph count accuracy, and reading order preservation. To support this training, we construct the Infinity-Doc-400K dataset, which we use to train Infinity-Parser, a vision-language model demonstrating robust generalization across various domains. Extensive evaluations on benchmarks including OmniDocBench, olmOCR-Bench, PubTabNet, and FinTabNet show that Infinity-Parser consistently achieves state-of-the-art performance across a broad range of document types, languages, and structural complexities, substantially outperforming both specialized document parsing systems and general-purpose vision-language models. We will release our code, dataset, and model to facilitate reproducible research in document parsing. |
| title | Infinity Parser: Layout Aware Reinforcement Learning for Scanned Document Parsing |
| topic | Computation and Language F.2.2; I.2.7 |
| url | https://arxiv.org/abs/2510.15349 |