@inproceedings{colonetti-benedet-etal-2026-challenges,
title = "Challenges in Image-Caption Association in {P}ortuguese: Evaluating the {CLIP} Model on the {FM}30{K} Dataset",
author = "Colonetti Benedet, Vit{\'o}ria and
Tamiosso, Gustavo Lopes and
Nunes, Rafael Oleques and
Balreira, Dennis Giovani",
editor = "Piperidis, Stelios and
Bel, N{\'u}ria and
van den Heuvel, Henk and
Ide, Nancy and
Krek, Simon and
Toral, Antonio",
booktitle = "Proceedings of the Fifteenth Language Resources and Evaluation Conference",
month = may,
year = "2026",
address = "Palma de Mallorca, Spain",
publisher = "ELRA Language Resource Association",
url = "https://aclanthology.org/2026.lrec-1.739/",
doi = "10.63317/5mwksx2sayjr",
pages = "9409--9419",
abstract = "In recent decades, multimodal models such as CLIP have achieved significant advances in associating images and texts. However, most of these advances stem from models trained almost exclusively in English, which limits their effectiveness in other languages. This challenge is particularly relevant for Brazilian Portuguese, a language that still lacks dedicated multimodal resources and relies predominantly on automatic translations. This work investigates the performance of CLIP-based multimodal models in the task of associating images and descriptions written in Brazilian Portuguese. The analysis begins with a zero-shot scenario, in which different CLIP variants are directly evaluated on the FM30k dataset, composed of images and captions originally written in Portuguese. An additional experiment with automatic translations is also conducted to examine the impact of language on cross-modal retrieval tasks. Subsequently, fine-tuning is performed on the textual encoder of the ViT-B/32 model, keeping the visual encoder frozen, with the goal of adapting the model to the target language. The results show that models originally trained in English perform worse in Portuguese, while linguistically adapted variants, either multilingual or Portuguese-specific, achieve superior performance. The proposed fine-tuning approach was able to reduce this performance gap, leading to notable improvements. In the image-to-text scenario, the model achieved an absolute increase of 27.65 percentage points in the Accuracy@1 metric, representing a 209{\%} relative gain over the original CLIP ViT-B/32. In the text-to-image scenario, the gain was 15.47 percentage points, amounting to an even higher 385{\%} relative improvement, contributing to a more balanced association between images and captions."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="colonetti-benedet-etal-2026-challenges">
<titleInfo>
<title>Challenges in Image-Caption Association in Portuguese: Evaluating the CLIP Model on the FM30K Dataset</title>
</titleInfo>
<name type="personal">
<namePart type="given">Vitória</namePart>
<namePart type="family">Colonetti Benedet</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Gustavo</namePart>
<namePart type="given">Lopes</namePart>
<namePart type="family">Tamiosso</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Rafael</namePart>
<namePart type="given">Oleques</namePart>
<namePart type="family">Nunes</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Dennis</namePart>
<namePart type="given">Giovani</namePart>
<namePart type="family">Balreira</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-05</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the Fifteenth Language Resources and Evaluation Conference</title>
</titleInfo>
<name type="personal">
<namePart type="given">Stelios</namePart>
<namePart type="family">Piperidis</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Núria</namePart>
<namePart type="family">Bel</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Henk</namePart>
<namePart type="family">van den Heuvel</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Nancy</namePart>
<namePart type="family">Ide</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Simon</namePart>
<namePart type="family">Krek</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Antonio</namePart>
<namePart type="family">Toral</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>ELRA Language Resource Association</publisher>
<place>
<placeTerm type="text">Palma de Mallorca, Spain</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>In recent decades, multimodal models such as CLIP have achieved significant advances in associating images and texts. However, most of these advances stem from models trained almost exclusively in English, which limits their effectiveness in other languages. This challenge is particularly relevant for Brazilian Portuguese, a language that still lacks dedicated multimodal resources and relies predominantly on automatic translations. This work investigates the performance of CLIP-based multimodal models in the task of associating images and descriptions written in Brazilian Portuguese. The analysis begins with a zero-shot scenario, in which different CLIP variants are directly evaluated on the FM30k dataset, composed of images and captions originally written in Portuguese. An additional experiment with automatic translations is also conducted to examine the impact of language on cross-modal retrieval tasks. Subsequently, fine-tuning is performed on the textual encoder of the ViT-B/32 model, keeping the visual encoder frozen, with the goal of adapting the model to the target language. The results show that models originally trained in English perform worse in Portuguese, while linguistically adapted variants, either multilingual or Portuguese-specific, achieve superior performance. The proposed fine-tuning approach was able to reduce this performance gap, leading to notable improvements. In the image-to-text scenario, the model achieved an absolute increase of 27.65 percentage points in the Accuracy@1 metric, representing a 209% relative gain over the original CLIP ViT-B/32. In the text-to-image scenario, the gain was 15.47 percentage points, amounting to an even higher 385% relative improvement, contributing to a more balanced association between images and captions.</abstract>
<identifier type="citekey">colonetti-benedet-etal-2026-challenges</identifier>
<identifier type="doi">10.63317/5mwksx2sayjr</identifier>
<location>
<url>https://aclanthology.org/2026.lrec-1.739/</url>
</location>
<part>
<date>2026-05</date>
<extent unit="page">
<start>9409</start>
<end>9419</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Challenges in Image-Caption Association in Portuguese: Evaluating the CLIP Model on the FM30K Dataset
%A Colonetti Benedet, Vitória
%A Tamiosso, Gustavo Lopes
%A Nunes, Rafael Oleques
%A Balreira, Dennis Giovani
%Y Piperidis, Stelios
%Y Bel, Núria
%Y van den Heuvel, Henk
%Y Ide, Nancy
%Y Krek, Simon
%Y Toral, Antonio
%S Proceedings of the Fifteenth Language Resources and Evaluation Conference
%D 2026
%8 May
%I ELRA Language Resource Association
%C Palma de Mallorca, Spain
%F colonetti-benedet-etal-2026-challenges
%X In recent decades, multimodal models such as CLIP have achieved significant advances in associating images and texts. However, most of these advances stem from models trained almost exclusively in English, which limits their effectiveness in other languages. This challenge is particularly relevant for Brazilian Portuguese, a language that still lacks dedicated multimodal resources and relies predominantly on automatic translations. This work investigates the performance of CLIP-based multimodal models in the task of associating images and descriptions written in Brazilian Portuguese. The analysis begins with a zero-shot scenario, in which different CLIP variants are directly evaluated on the FM30k dataset, composed of images and captions originally written in Portuguese. An additional experiment with automatic translations is also conducted to examine the impact of language on cross-modal retrieval tasks. Subsequently, fine-tuning is performed on the textual encoder of the ViT-B/32 model, keeping the visual encoder frozen, with the goal of adapting the model to the target language. The results show that models originally trained in English perform worse in Portuguese, while linguistically adapted variants, either multilingual or Portuguese-specific, achieve superior performance. The proposed fine-tuning approach was able to reduce this performance gap, leading to notable improvements. In the image-to-text scenario, the model achieved an absolute increase of 27.65 percentage points in the Accuracy@1 metric, representing a 209% relative gain over the original CLIP ViT-B/32. In the text-to-image scenario, the gain was 15.47 percentage points, amounting to an even higher 385% relative improvement, contributing to a more balanced association between images and captions.
%R 10.63317/5mwksx2sayjr
%U https://aclanthology.org/2026.lrec-1.739/
%U https://doi.org/10.63317/5mwksx2sayjr
%P 9409-9419
Markdown (Informal)
[Challenges in Image-Caption Association in Portuguese: Evaluating the CLIP Model on the FM30K Dataset](https://aclanthology.org/2026.lrec-1.739/) (Colonetti Benedet et al., LREC 2026)
ACL