Thinking with Anchors: Grounded and Efficient Document Reasoning
Quick Answer
This paper shows that ADOPD 2026 enhances document understanding by integrating region semantics, spatial relations, and visual structure into a unified framework.
Quick Take
It introduces a shared vocabulary of visual anchors, improving semantic tagging and enabling structured outputs for downstream reasoning, addressing challenges in dense counting tasks. This approach marks a significant shift from traditional localization to anchor-grounded document intelligence.
Key Points
- ADOPD 2026 extends ADOPD 2024 with human-cleaned captions and semantic tags.
- It transforms detection outputs into structured anchors for improved reasoning.
- The model addresses long-tail semantic failures often hidden in standard benchmarks.
- Current models struggle with dense counting tasks, highlighting the need for this framework.
- This approach shifts document understanding from localization to anchor-grounded intelligence.
Paper Resources
📖 Reader Mode
~2 min readAuthors:Sichen Zhu, Yuchen Zhu, Wenzhuo Xu, Jason Kuen, Wanrong Zhu, Jing Shi, Xuan Shen, Quanyi Wang, Yiwei Wang, Yujun Cai, Bing Shuai, Qin Zhang, Yongxin Chen, Shilong Liu, Molei Tao, Jiuxiang Gu
Abstract:Existing document understanding benchmarks have largely focused on locating page elements, yet real-world document intelligence requires models to reason jointly about region semantics, spatial relations, and visual structure. We present ADOPD 2026, a reasoning-oriented extension of ADOPD that turns page decomposition into spatially grounded document understanding. ADOPD 2026 enriches page anchors inherited from ADOPD 2024 dataset with human-cleaned captions, semantic tags, and generated chain-of-thought (CoT) traces grounded to document regions. Instead of treating boxes, masks, and tags as independent supervision signals, we cast text blocks, visual entities, semantic labels, bounding boxes, and polygon masks as a shared vocabulary of visual anchors. This representation supports three connected capabilities. First, region-level semantic tagging asks models to identify document element types from both page context and local appearance, revealing long-tail semantic failures that standard layout benchmarks often hide. Second, unified vision-language grounding generates text regions and visual entities together with coordinates or polygonal outlines, transforming detection and segmentation outputs into structured anchors that can be reused by downstream reasoning systems. Third, current state-of-the-art models still struggle with dense counting tasks evaluated on DocCount, a benchmark derived from ADOPD 2026, highlighting the need for the Thinking-with-Anchors pipeline in document semantic understanding. By connecting page decomposition to verifiable visual-anchor reasoning, ADOPD 2026 provides a task framework that moves document understanding beyond localization toward anchor-grounded document intelligence.
| Subjects: | Computer Vision and Pattern Recognition (cs.CV) |
| Cite as: | arXiv:2608.04424 [cs.CV] |
| (or arXiv:2608.04424v1 [cs.CV] for this version) | |
| https://doi.org/10.48550/arXiv.2608.04424 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Sichen Zhu [view email]
[v1]
Wed, 5 Aug 2026 04:09:05 UTC (28,901 KB)
— Originally published at arxiv.org
Want this in your inbox every morning?
Daily brief at your local 8am — bilingual EN/中文, free.
More from arXiv cs.CV
See more →ProMoE-FL: Prototype-conditioned Mixture of Experts for Multimodal Federated Learning with Missing Modalities
ProMoE-FL introduces a Prototype-conditioned Mixture-of-Experts framework for multimodal federated learning, effectively addressing missing modalities. It outperforms existing methods on four chest X-ray datasets, demonstrating superior feature synthesis capabilities in both homogeneous and heterogeneous settings.