StrucTab: A Structured Optimization Framework for Table Parsing

2026-06-29

AI Summary

This work addresses the challenges in table image parsing arising from the absence of explicit intermediate reasoning and unstable reward design, proposing a unified reinforcement learning framework named Uni-TabRL. The approach decomposes the parsing task into subtasks such as row/column counting and merged cell detection, progressively integrating them through a sequential reasoning strategy. It introduces a decomposed reward mechanism grounded in validity, structure, and content, combined with structured intermediate supervision and visual-language model guidance for optimization. Concurrently, the authors release TableVerse-5K, a large-scale benchmark of real-world table images. Experiments demonstrate that Uni-TabRL achieves state-of-the-art performance across all public benchmarks and significantly outperforms existing methods on TableVerse-5K, confirming its effectiveness and strong generalization capability.

Abstract

Table parsing aims to convert table images into structured, machine-readable representations, a task requiring the joint perception of complex spatial layouts and textual content. While recent vision-language models (VLMs) enable end-to-end parsing, they typically rely on direct supervision of the final output, thereby bypassing the explicit intermediate reasoning that is crucial for understanding complex table structures. Furthermore, attempts to optimize these models using reinforcement learning (RL) are often hindered by unstable or ambiguous reward designs, limiting potential performance gains. To address these limitations, we propose StrucTab, a table parsing model learned through intermediate structural supervision and reward decomposition. At the modeling level, by decomposing the parsing process into human-inspired subtasks, such as row-column counting and merged-cell analysis, StrucTab progressively unifies them through a sequential reasoning strategy. At the optimization level, we introduce Uni-TabRL, a unified RL framework that leverages decomposed rewards (validity, structure, and content) to provide stable and informative optimization signals. Finally, at the evaluation level, we present TableVerse-5K, a large-scale, challenging benchmark encompassing diverse, real-world table scenarios. Extensive experiments demonstrate the state-of-the-art performance of StrucTab across all evaluated public benchmarks and significant improvements on TableVerse-5K, validating the effectiveness of explicit structural modeling and decomposed reward optimization. Code and benchmark are publicly available at https://github.com/VirtualLUOUCAS/StrucTab.

Problem

• table parsing
• structured representation
• vision-language models
• reinforcement learning
• reward design

Innovation

• structured supervision
• reward decomposition
• table parsing
• vision-language models
• reinforcement learning