@inproceedings{nunes-etal-2025-benchmarking,
title = "Benchmarking Table Extraction: Multimodal {LLM}s vs Traditional {OCR}",
author = "Nunes, Guilherme and
Rolla, Vitor and
Pereira, Duarte and
Alves, Vasco and
Carreiro, Andre and
Baptista, M{\'a}rcia",
editor = "Fei, Hao and
Tu, Kewei and
Zhang, Yuhui and
Hu, Xiang and
Han, Wenjuan and
Jia, Zixia and
Zheng, Zilong and
Cao, Yixin and
Zhang, Meishan and
Lu, Wei and
Siddharth, N. and
{\O}vrelid, Lilja and
Xue, Nianwen and
Zhang, Yue",
booktitle = "Proceedings of the 1st Joint Workshop on Large Language Models and Structure Modeling (XLLM 2025)",
month = aug,
year = "2025",
address = "Vienna, Austria",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2025.xllm-1.2/",
doi = "10.18653/v1/2025.xllm-1.2",
pages = "8--15",
ISBN = "979-8-89176-286-2",
abstract = "This paper compares two approaches for table extraction from images: deep learning computer vision and Multimodal Large Language Models (MLLMs). Computer vision models for table extraction, such as the Table Transformer model (TATR), have enhanced the extraction of complex table structural layouts by leveraging deep learning for precise structural recognition combined with traditional Optical Character Recognition (OCR). Conversely, MLLMs, which process both text and image inputs, present a novel approach by potentially bypassing the limitations of TATR plus OCR methods altogether. Models such as GPT-4o, Phi-3 Vision, and Granite Vision 3.2 demonstrate the potential of MLLMs to analyze and interpret table images directly, offering enhanced accuracy and robust extraction capabilities. A state-of-the-art metric like Grid Table Similarity (GriTS) evaluated these methodologies, providing nuanced insights into structural and text content effectiveness. Utilizing the PubTables-1M dataset, a comprehensive and widely used benchmark in the field, this study highlights the strengths and limitations of each approach, setting the stage for future innovations in table extraction technologies. Deep learning computer vision techniques still have a slight edge when extracting table structural layout, but in terms of text cell content, MLLMs are far better."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="nunes-etal-2025-benchmarking">
<titleInfo>
<title>Benchmarking Table Extraction: Multimodal LLMs vs Traditional OCR</title>
</titleInfo>
<name type="personal">
<namePart type="given">Guilherme</namePart>
<namePart type="family">Nunes</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Vitor</namePart>
<namePart type="family">Rolla</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Duarte</namePart>
<namePart type="family">Pereira</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Vasco</namePart>
<namePart type="family">Alves</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Andre</namePart>
<namePart type="family">Carreiro</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Márcia</namePart>
<namePart type="family">Baptista</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2025-08</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 1st Joint Workshop on Large Language Models and Structure Modeling (XLLM 2025)</title>
</titleInfo>
<name type="personal">
<namePart type="given">Hao</namePart>
<namePart type="family">Fei</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Kewei</namePart>
<namePart type="family">Tu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yuhui</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xiang</namePart>
<namePart type="family">Hu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Wenjuan</namePart>
<namePart type="family">Han</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zixia</namePart>
<namePart type="family">Jia</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zilong</namePart>
<namePart type="family">Zheng</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yixin</namePart>
<namePart type="family">Cao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Meishan</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Wei</namePart>
<namePart type="family">Lu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">N</namePart>
<namePart type="family">Siddharth</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Lilja</namePart>
<namePart type="family">Øvrelid</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Nianwen</namePart>
<namePart type="family">Xue</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yue</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Vienna, Austria</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-286-2</identifier>
</relatedItem>
<abstract>This paper compares two approaches for table extraction from images: deep learning computer vision and Multimodal Large Language Models (MLLMs). Computer vision models for table extraction, such as the Table Transformer model (TATR), have enhanced the extraction of complex table structural layouts by leveraging deep learning for precise structural recognition combined with traditional Optical Character Recognition (OCR). Conversely, MLLMs, which process both text and image inputs, present a novel approach by potentially bypassing the limitations of TATR plus OCR methods altogether. Models such as GPT-4o, Phi-3 Vision, and Granite Vision 3.2 demonstrate the potential of MLLMs to analyze and interpret table images directly, offering enhanced accuracy and robust extraction capabilities. A state-of-the-art metric like Grid Table Similarity (GriTS) evaluated these methodologies, providing nuanced insights into structural and text content effectiveness. Utilizing the PubTables-1M dataset, a comprehensive and widely used benchmark in the field, this study highlights the strengths and limitations of each approach, setting the stage for future innovations in table extraction technologies. Deep learning computer vision techniques still have a slight edge when extracting table structural layout, but in terms of text cell content, MLLMs are far better.</abstract>
<identifier type="citekey">nunes-etal-2025-benchmarking</identifier>
<identifier type="doi">10.18653/v1/2025.xllm-1.2</identifier>
<location>
<url>https://aclanthology.org/2025.xllm-1.2/</url>
</location>
<part>
<date>2025-08</date>
<extent unit="page">
<start>8</start>
<end>15</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Benchmarking Table Extraction: Multimodal LLMs vs Traditional OCR
%A Nunes, Guilherme
%A Rolla, Vitor
%A Pereira, Duarte
%A Alves, Vasco
%A Carreiro, Andre
%A Baptista, Márcia
%Y Fei, Hao
%Y Tu, Kewei
%Y Zhang, Yuhui
%Y Hu, Xiang
%Y Han, Wenjuan
%Y Jia, Zixia
%Y Zheng, Zilong
%Y Cao, Yixin
%Y Zhang, Meishan
%Y Lu, Wei
%Y Siddharth, N.
%Y Øvrelid, Lilja
%Y Xue, Nianwen
%Y Zhang, Yue
%S Proceedings of the 1st Joint Workshop on Large Language Models and Structure Modeling (XLLM 2025)
%D 2025
%8 August
%I Association for Computational Linguistics
%C Vienna, Austria
%@ 979-8-89176-286-2
%F nunes-etal-2025-benchmarking
%X This paper compares two approaches for table extraction from images: deep learning computer vision and Multimodal Large Language Models (MLLMs). Computer vision models for table extraction, such as the Table Transformer model (TATR), have enhanced the extraction of complex table structural layouts by leveraging deep learning for precise structural recognition combined with traditional Optical Character Recognition (OCR). Conversely, MLLMs, which process both text and image inputs, present a novel approach by potentially bypassing the limitations of TATR plus OCR methods altogether. Models such as GPT-4o, Phi-3 Vision, and Granite Vision 3.2 demonstrate the potential of MLLMs to analyze and interpret table images directly, offering enhanced accuracy and robust extraction capabilities. A state-of-the-art metric like Grid Table Similarity (GriTS) evaluated these methodologies, providing nuanced insights into structural and text content effectiveness. Utilizing the PubTables-1M dataset, a comprehensive and widely used benchmark in the field, this study highlights the strengths and limitations of each approach, setting the stage for future innovations in table extraction technologies. Deep learning computer vision techniques still have a slight edge when extracting table structural layout, but in terms of text cell content, MLLMs are far better.
%R 10.18653/v1/2025.xllm-1.2
%U https://aclanthology.org/2025.xllm-1.2/
%U https://doi.org/10.18653/v1/2025.xllm-1.2
%P 8-15
Markdown (Informal)
[Benchmarking Table Extraction: Multimodal LLMs vs Traditional OCR](https://aclanthology.org/2025.xllm-1.2/) (Nunes et al., XLLM 2025)
ACL
- Guilherme Nunes, Vitor Rolla, Duarte Pereira, Vasco Alves, Andre Carreiro, and Márcia Baptista. 2025. Benchmarking Table Extraction: Multimodal LLMs vs Traditional OCR. In Proceedings of the 1st Joint Workshop on Large Language Models and Structure Modeling (XLLM 2025), pages 8–15, Vienna, Austria. Association for Computational Linguistics.