@inproceedings{yan-etal-2025-fiha,
title = "{FIHA}: Automated Fine-grained Hallucinations Evaluations in Large Vision Language Models with {D}avidson Scene Graphs",
author = "Yan, Bowen and
Zhang, Zhengsong and
Jing, Liqiang and
Hossain, Eftekhar and
Du, Xinya",
editor = "Che, Wanxiang and
Nabende, Joyce and
Shutova, Ekaterina and
Pilehvar, Mohammad Taher",
booktitle = "Findings of the Association for Computational Linguistics: ACL 2025",
month = jul,
year = "2025",
address = "Vienna, Austria",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2025.findings-acl.622/",
doi = "10.18653/v1/2025.findings-acl.622",
pages = "12014--12026",
ISBN = "979-8-89176-256-5",
abstract = "The rapid development of Large Vision-Language Models (LVLMs) often comes with widespread hallucination issues, making cost-effective and comprehensive assessments increasingly vital. Current approaches mainly rely on costly annotations and are not comprehensive {--} in terms of evaluating all aspects, such as relations, attributes, and dependencies between aspects. Therefore, we introduce the FIHA (automated Fine-graIned Hallucination evAluation in LVLMs), which could access LVLMs hallucination in an LLM-free and annotation-free way and model the dependency between different types of hallucinations. FIHA can generate Q{\&}A pairs on any image dataset at minimal cost, enabling hallucination assessment from both image and caption. Based on this approach, we introduce a benchmark called FIHA-v1, which consists of diverse questions on various images from three datasets. Furthermore, we use the Davidson Scene Graph (DSG) to organize the structure among Q{\&}A pairs, in which we can increase the reliability of the evaluation. We evaluate representative models using FIHA-v1, highlighting their limitations and challenges. We released our code and data at https://github.com/confidentzzzs/FIHA."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="yan-etal-2025-fiha">
<titleInfo>
<title>FIHA: Automated Fine-grained Hallucinations Evaluations in Large Vision Language Models with Davidson Scene Graphs</title>
</titleInfo>
<name type="personal">
<namePart type="given">Bowen</namePart>
<namePart type="family">Yan</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zhengsong</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Liqiang</namePart>
<namePart type="family">Jing</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Eftekhar</namePart>
<namePart type="family">Hossain</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xinya</namePart>
<namePart type="family">Du</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2025-07</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: ACL 2025</title>
</titleInfo>
<name type="personal">
<namePart type="given">Wanxiang</namePart>
<namePart type="family">Che</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Joyce</namePart>
<namePart type="family">Nabende</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ekaterina</namePart>
<namePart type="family">Shutova</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Mohammad</namePart>
<namePart type="given">Taher</namePart>
<namePart type="family">Pilehvar</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">979-8-89176-256-5</identifier>
</relatedItem>
<abstract>The rapid development of Large Vision-Language Models (LVLMs) often comes with widespread hallucination issues, making cost-effective and comprehensive assessments increasingly vital. Current approaches mainly rely on costly annotations and are not comprehensive – in terms of evaluating all aspects, such as relations, attributes, and dependencies between aspects. Therefore, we introduce the FIHA (automated Fine-graIned Hallucination evAluation in LVLMs), which could access LVLMs hallucination in an LLM-free and annotation-free way and model the dependency between different types of hallucinations. FIHA can generate Q&A pairs on any image dataset at minimal cost, enabling hallucination assessment from both image and caption. Based on this approach, we introduce a benchmark called FIHA-v1, which consists of diverse questions on various images from three datasets. Furthermore, we use the Davidson Scene Graph (DSG) to organize the structure among Q&A pairs, in which we can increase the reliability of the evaluation. We evaluate representative models using FIHA-v1, highlighting their limitations and challenges. We released our code and data at https://github.com/confidentzzzs/FIHA.</abstract>
<identifier type="citekey">yan-etal-2025-fiha</identifier>
<identifier type="doi">10.18653/v1/2025.findings-acl.622</identifier>
<location>
<url>https://aclanthology.org/2025.findings-acl.622/</url>
</location>
<part>
<date>2025-07</date>
<extent unit="page">
<start>12014</start>
<end>12026</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T FIHA: Automated Fine-grained Hallucinations Evaluations in Large Vision Language Models with Davidson Scene Graphs
%A Yan, Bowen
%A Zhang, Zhengsong
%A Jing, Liqiang
%A Hossain, Eftekhar
%A Du, Xinya
%Y Che, Wanxiang
%Y Nabende, Joyce
%Y Shutova, Ekaterina
%Y Pilehvar, Mohammad Taher
%S Findings of the Association for Computational Linguistics: ACL 2025
%D 2025
%8 July
%I Association for Computational Linguistics
%C Vienna, Austria
%@ 979-8-89176-256-5
%F yan-etal-2025-fiha
%X The rapid development of Large Vision-Language Models (LVLMs) often comes with widespread hallucination issues, making cost-effective and comprehensive assessments increasingly vital. Current approaches mainly rely on costly annotations and are not comprehensive – in terms of evaluating all aspects, such as relations, attributes, and dependencies between aspects. Therefore, we introduce the FIHA (automated Fine-graIned Hallucination evAluation in LVLMs), which could access LVLMs hallucination in an LLM-free and annotation-free way and model the dependency between different types of hallucinations. FIHA can generate Q&A pairs on any image dataset at minimal cost, enabling hallucination assessment from both image and caption. Based on this approach, we introduce a benchmark called FIHA-v1, which consists of diverse questions on various images from three datasets. Furthermore, we use the Davidson Scene Graph (DSG) to organize the structure among Q&A pairs, in which we can increase the reliability of the evaluation. We evaluate representative models using FIHA-v1, highlighting their limitations and challenges. We released our code and data at https://github.com/confidentzzzs/FIHA.
%R 10.18653/v1/2025.findings-acl.622
%U https://aclanthology.org/2025.findings-acl.622/
%U https://doi.org/10.18653/v1/2025.findings-acl.622
%P 12014-12026
Markdown (Informal)
[FIHA: Automated Fine-grained Hallucinations Evaluations in Large Vision Language Models with Davidson Scene Graphs](https://aclanthology.org/2025.findings-acl.622/) (Yan et al., Findings 2025)
ACL