Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2605.27978 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866910264718262272 |
|---|---|
| author | Jiang, Kaitao Gong, Ruiyan Cheng, Xiaolong Niu, Kangning Li, Tianlun Xu, Mu |
| author_facet | Jiang, Kaitao Gong, Ruiyan Cheng, Xiaolong Niu, Kangning Li, Tianlun Xu, Mu |
| contents | We introduce ABot-OCR, an end-to-end vision-language model that transcribes a page image directly into clean Markdown in a single forward pass. By doing so, our approach completely eliminates the need for brittle modular orchestration. To maximize parsing fidelity, we develop a dedicated data engine to provide large-scale, structurally consistent supervision. Furthermore, we propose Decoupled Heterogeneous Document Optimization, a structure-constrained reinforcement learning method that sharpens textual accuracy and strictly enforces markup well-formedness beyond supervised fine-tuning alone. Extensive evaluations demonstrate the superior performance of our framework. On the OmniDocBench v1.5 and v1.6 benchmarks, ABot-OCR achieves state-of-the-art scores of 92.81 and 93.30 among all end-to-end systems, substantially narrowing the performance gap relative to strong pipeline baselines. Finally, comprehensive multilingual text recognition across ten diverse languages further confirms the robust generalizability of ABot-OCR. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2605_27978 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | ABot-OCR Technical Report Jiang, Kaitao Gong, Ruiyan Cheng, Xiaolong Niu, Kangning Li, Tianlun Xu, Mu Computer Vision and Pattern Recognition We introduce ABot-OCR, an end-to-end vision-language model that transcribes a page image directly into clean Markdown in a single forward pass. By doing so, our approach completely eliminates the need for brittle modular orchestration. To maximize parsing fidelity, we develop a dedicated data engine to provide large-scale, structurally consistent supervision. Furthermore, we propose Decoupled Heterogeneous Document Optimization, a structure-constrained reinforcement learning method that sharpens textual accuracy and strictly enforces markup well-formedness beyond supervised fine-tuning alone. Extensive evaluations demonstrate the superior performance of our framework. On the OmniDocBench v1.5 and v1.6 benchmarks, ABot-OCR achieves state-of-the-art scores of 92.81 and 93.30 among all end-to-end systems, substantially narrowing the performance gap relative to strong pipeline baselines. Finally, comprehensive multilingual text recognition across ten diverse languages further confirms the robust generalizability of ABot-OCR. |
| title | ABot-OCR Technical Report |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2605.27978 |