@inproceedings{luu-etal-2026-speci,
title = "{SPECI}: Fine-grained Evaluation of Specificity for Image Captioning",
author = "Luu, Son T. and
Nguyen, Hiep K. V. and
Bui, Trung and
Lai, Viet Dac and
Nguyen, Le-Minh",
editor = "Zhang, Yue and
Xu, Ruifeng and
Purwarianti, Ayu and
Rong, Tong and
Suzuki, Jun",
booktitle = "Proceedings of the 15th International Joint Conference on Natural Language Processing and the 5th Conference of the {A}sia-Pacific Chapter of the {A}ssociation for {C}omputational {L}inguistics (Volume 1: Long Papers)",
month = nov,
year = "2026",
address = "Hengqin, China",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.aacl-long.7/",
pages = "140--156",
ISBN = "979-8-89176-439-2",
abstract = "Image Captioning lies at the center of synthetic data generation in computer vision. The quality of the caption defines the behavior of the resulting model trained on image-caption pairs generated by Image Captioning. Measuring the specificity of image captioning plays an important role in supplying high-quality captions for the downstream tasks. To address the challenge of automatically assessing the quality of the caption, we propose SPECI, an automatic metric that decomposes dense captions into atomic facts and evaluates specificity across visual criteria such as granularity, spatial detail, color, and linguistic focus using WordNet-based hyponym levels. Empirical results on the benchmark datasets demonstrate significant correlation with human judgments, and its application in image captioning and image reconstruction tasks highlights the practical utility of SPECI. The source code of SPECI is available at https://github.com/sonlam1102/SPECI."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="luu-etal-2026-speci">
<titleInfo>
<title>SPECI: Fine-grained Evaluation of Specificity for Image Captioning</title>
</titleInfo>
<name type="personal">
<namePart type="given">Son</namePart>
<namePart type="given">T</namePart>
<namePart type="family">Luu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Hiep</namePart>
<namePart type="given">K</namePart>
<namePart type="given">V</namePart>
<namePart type="family">Nguyen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Trung</namePart>
<namePart type="family">Bui</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Viet</namePart>
<namePart type="given">Dac</namePart>
<namePart type="family">Lai</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Le-Minh</namePart>
<namePart type="family">Nguyen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-11</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 15th International Joint Conference on Natural Language Processing and the 5th Conference of the Asia-Pacific Chapter of the Association for Computational Linguistics (Volume 1: Long Papers)</title>
</titleInfo>
<name type="personal">
<namePart type="given">Yue</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ruifeng</namePart>
<namePart type="family">Xu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ayu</namePart>
<namePart type="family">Purwarianti</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Tong</namePart>
<namePart type="family">Rong</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jun</namePart>
<namePart type="family">Suzuki</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Hengqin, China</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-439-2</identifier>
</relatedItem>
<abstract>Image Captioning lies at the center of synthetic data generation in computer vision. The quality of the caption defines the behavior of the resulting model trained on image-caption pairs generated by Image Captioning. Measuring the specificity of image captioning plays an important role in supplying high-quality captions for the downstream tasks. To address the challenge of automatically assessing the quality of the caption, we propose SPECI, an automatic metric that decomposes dense captions into atomic facts and evaluates specificity across visual criteria such as granularity, spatial detail, color, and linguistic focus using WordNet-based hyponym levels. Empirical results on the benchmark datasets demonstrate significant correlation with human judgments, and its application in image captioning and image reconstruction tasks highlights the practical utility of SPECI. The source code of SPECI is available at https://github.com/sonlam1102/SPECI.</abstract>
<identifier type="citekey">luu-etal-2026-speci</identifier>
<location>
<url>https://aclanthology.org/2026.aacl-long.7/</url>
</location>
<part>
<date>2026-11</date>
<extent unit="page">
<start>140</start>
<end>156</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T SPECI: Fine-grained Evaluation of Specificity for Image Captioning
%A Luu, Son T.
%A Nguyen, Hiep K. V.
%A Bui, Trung
%A Lai, Viet Dac
%A Nguyen, Le-Minh
%Y Zhang, Yue
%Y Xu, Ruifeng
%Y Purwarianti, Ayu
%Y Rong, Tong
%Y Suzuki, Jun
%S Proceedings of the 15th International Joint Conference on Natural Language Processing and the 5th Conference of the Asia-Pacific Chapter of the Association for Computational Linguistics (Volume 1: Long Papers)
%D 2026
%8 November
%I Association for Computational Linguistics
%C Hengqin, China
%@ 979-8-89176-439-2
%F luu-etal-2026-speci
%X Image Captioning lies at the center of synthetic data generation in computer vision. The quality of the caption defines the behavior of the resulting model trained on image-caption pairs generated by Image Captioning. Measuring the specificity of image captioning plays an important role in supplying high-quality captions for the downstream tasks. To address the challenge of automatically assessing the quality of the caption, we propose SPECI, an automatic metric that decomposes dense captions into atomic facts and evaluates specificity across visual criteria such as granularity, spatial detail, color, and linguistic focus using WordNet-based hyponym levels. Empirical results on the benchmark datasets demonstrate significant correlation with human judgments, and its application in image captioning and image reconstruction tasks highlights the practical utility of SPECI. The source code of SPECI is available at https://github.com/sonlam1102/SPECI.
%U https://aclanthology.org/2026.aacl-long.7/
%P 140-156
Markdown (Informal)
[SPECI: Fine-grained Evaluation of Specificity for Image Captioning](https://aclanthology.org/2026.aacl-long.7/) (Luu et al., AACL-IJCNLP 2026)
ACL
- Son T. Luu, Hiep K. V. Nguyen, Trung Bui, Viet Dac Lai, and Le-Minh Nguyen. 2026. SPECI: Fine-grained Evaluation of Specificity for Image Captioning. In Proceedings of the 15th International Joint Conference on Natural Language Processing and the 5th Conference of the Asia-Pacific Chapter of the Association for Computational Linguistics (Volume 1: Long Papers), pages 140–156, Hengqin, China. Association for Computational Linguistics.