Table Structure Recognition

Follow this topic Ask about this topic

Table structure recognition studies how to recover a table’s rows, columns, cell boundaries, spanning cells, and hierarchy from images or document pages, often alongside cell text. Approaches include grid and object detection, graph-based prediction, autoregressive HTML or OTSL generation, and vision-language models guided by spatial or structural priors. Systems are commonly evaluated on benchmarks such as PubTabNet, FinTabNet, and PubTables-1M using structural and content metrics such as TEDS and GriTS. A central challenge is handling complex, degraded, or unfamiliar tables: strong scores on clean benchmarks do not guarantee reliable real-world extraction, and accurate structure alone does not ensure sound question answering or arithmetic.

Latest papers

32 papers1 paper in the last four weeks, down 83% on the four weeks before. 0.0% of all new papers.

Orthogonal Hierarchical Decomposition for Structure-Aware Table Understanding with Large Language Models
Feb 2, 2026Bin Cao, Huixian Lu, Chenwen Ma +3Table QA Table Structure Recognition
Zhejiang University of Technology, China · Zhejiang Key Laboratory of Visual Information Intelligent Processing, China · University of Aberdeen, UK +1
Enhancing Table Recognition with Vision LLMs: A Benchmark and Neighbor-Guided Toolchain Reasoner
Dec 30, 2024Yitong Zhou, Mingyue Cheng, Qingyang Mao +3Table Structure Recognition VLM Evaluation
State Key Laboratory of Cognitive Intelligence, University of Science and Technology of China · Artificial Intelligence Research Institute, iFLYTEK Co., Ltd

Feb 2, 2026J/K move · Enter open · S save

Orthogonal Hierarchical Decomposition for Structure-Aware Table Understanding with Large Language Models

Bin Cao, Huixian Lu, Chenwen Ma, Ting Wang+2

Zhejiang University of Technology, China · Zhejiang Key Laboratory of Visual Information Intelligent Processing, China · University of Aberdeen, UK+1

Table QA Table Structure Recognition Hierarchical Representation Learning

Open paper AskPDF arXiv

Complex tables with multi-level headers, merged cells and heterogeneous layouts pose persistent challenges for LLMs in both understanding and reasoning. Existing approaches typically rely on table linearization or normalized grid modeling. However, these representations struggle to explicitly capture hierarchical structures and cross-dimensional dependencies, which can lead to misalignment between structural semantics and textual representations for non-standard tables. To address this issue, we propose an Orthogonal Hierarchical Decomposition (OHD) framework that constructs structure-preserving input representations of complex tables for LLMs. OHD introduces an Orthogonal Tree Induction (OTI) method based on spatial--semantic co-constraints, which decomposes irregular tables into a column tree and a row tree to capture vertical and horizontal hierarchical dependencies, respectively. Building on this representation, we design a dual-pathway association protocol to symmetrically reconstruct semantic lineage of each cell, and incorporate an LLM as a semantic arbitrator to align multi-level semantic information. We evaluate OHD framework on two complex table question answering benchmarks, AITQA and HiTab. Experimental results show that OHD consistently outperforms existing representation paradigms across multiple evaluation metrics.

Latest papers

Orthogonal Hierarchical Decomposition for Structure-Aware Table Understanding with Large Language Models

Enhancing Table Recognition with Vision LLMs: A Benchmark and Neighbor-Guided Toolchain Reasoner