@inproceedings{soumik-etal-2026-edits,
title = "Edits in the Wild: A Multilingual Benchmark for Localizing On-Device Generative Image Edits on Consumer Platforms",
author = "Soumik, Mohd. Farhan Israk and
Fime, Awal Ahmed and
Kankanamge, Malithi Mithsara Wanniarachchi and
Aryal, Arun Babu and
Haghighi, Hoda Joudi and
Koirala, Bibek and
Imteaj, Ahmed and
Shahid, Abdur R.",
editor = "Lee, Hung-Yi and
Dev, Sunipa and
Li, Jessy and
Elliott, Desmond",
booktitle = "Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.emnlp-main.1661/",
pages = "37756--37789",
ISBN = "979-8-89176-459-0",
abstract = "Consumer devices now ship with generative image editing tools that can remove objects, move subjects, and create new visual content directly on phones, tablets, and laptops. Yet current benchmarks largely miss this setting: real on-device edits produced by black-box consumer generative AI editing tools, with variation across platforms, devices, edit operations, and post-edit compression. We introduce \textsc{ConsumerGenEdit}, a large-scale multilingual benchmark for language-guided localization of on-device generative image edits. The dataset contains 4,000 royalty-free source images and 48,000 manually edited images created on Apple, Google, and Samsung devices, covering object removal, object movement, and sketch-based object generation. To study robustness under realistic sharing conditions, we add 144,000 learned-compression variants at three quality levels. Each example includes platform, device, editing operation, compression metadata, bounding-box annotations, and edit descriptions in English, Bangla, and Persian/Farsi generated with five vision-language models (VLMs). We evaluate five separate benchmark VLMs on edit detection and edit-operation classification. By focusing on where and how consumer AI edits occur, rather than only whether an image is synthetic, \textsc{ConsumerGenEdit} provides a practical testbed for multilingual VLM grounding, image provenance, and robust interpretation of AI-edited media."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="soumik-etal-2026-edits">
<titleInfo>
<title>Edits in the Wild: A Multilingual Benchmark for Localizing On-Device Generative Image Edits on Consumer Platforms</title>
</titleInfo>
<name type="personal">
<namePart type="given">Mohd.</namePart>
<namePart type="given">Farhan</namePart>
<namePart type="given">Israk</namePart>
<namePart type="family">Soumik</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Awal</namePart>
<namePart type="given">Ahmed</namePart>
<namePart type="family">Fime</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Malithi</namePart>
<namePart type="given">Mithsara</namePart>
<namePart type="given">Wanniarachchi</namePart>
<namePart type="family">Kankanamge</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Arun</namePart>
<namePart type="given">Babu</namePart>
<namePart type="family">Aryal</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Hoda</namePart>
<namePart type="given">Joudi</namePart>
<namePart type="family">Haghighi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Bibek</namePart>
<namePart type="family">Koirala</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ahmed</namePart>
<namePart type="family">Imteaj</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Abdur</namePart>
<namePart type="given">R</namePart>
<namePart type="family">Shahid</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-10</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing</title>
</titleInfo>
<name type="personal">
<namePart type="given">Hung-Yi</namePart>
<namePart type="family">Lee</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sunipa</namePart>
<namePart type="family">Dev</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jessy</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Desmond</namePart>
<namePart type="family">Elliott</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Budapest, Hungary</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-459-0</identifier>
</relatedItem>
<abstract>Consumer devices now ship with generative image editing tools that can remove objects, move subjects, and create new visual content directly on phones, tablets, and laptops. Yet current benchmarks largely miss this setting: real on-device edits produced by black-box consumer generative AI editing tools, with variation across platforms, devices, edit operations, and post-edit compression. We introduce ConsumerGenEdit, a large-scale multilingual benchmark for language-guided localization of on-device generative image edits. The dataset contains 4,000 royalty-free source images and 48,000 manually edited images created on Apple, Google, and Samsung devices, covering object removal, object movement, and sketch-based object generation. To study robustness under realistic sharing conditions, we add 144,000 learned-compression variants at three quality levels. Each example includes platform, device, editing operation, compression metadata, bounding-box annotations, and edit descriptions in English, Bangla, and Persian/Farsi generated with five vision-language models (VLMs). We evaluate five separate benchmark VLMs on edit detection and edit-operation classification. By focusing on where and how consumer AI edits occur, rather than only whether an image is synthetic, ConsumerGenEdit provides a practical testbed for multilingual VLM grounding, image provenance, and robust interpretation of AI-edited media.</abstract>
<identifier type="citekey">soumik-etal-2026-edits</identifier>
<location>
<url>https://aclanthology.org/2026.emnlp-main.1661/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>37756</start>
<end>37789</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Edits in the Wild: A Multilingual Benchmark for Localizing On-Device Generative Image Edits on Consumer Platforms
%A Soumik, Mohd. Farhan Israk
%A Fime, Awal Ahmed
%A Kankanamge, Malithi Mithsara Wanniarachchi
%A Aryal, Arun Babu
%A Haghighi, Hoda Joudi
%A Koirala, Bibek
%A Imteaj, Ahmed
%A Shahid, Abdur R.
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-459-0
%F soumik-etal-2026-edits
%X Consumer devices now ship with generative image editing tools that can remove objects, move subjects, and create new visual content directly on phones, tablets, and laptops. Yet current benchmarks largely miss this setting: real on-device edits produced by black-box consumer generative AI editing tools, with variation across platforms, devices, edit operations, and post-edit compression. We introduce ConsumerGenEdit, a large-scale multilingual benchmark for language-guided localization of on-device generative image edits. The dataset contains 4,000 royalty-free source images and 48,000 manually edited images created on Apple, Google, and Samsung devices, covering object removal, object movement, and sketch-based object generation. To study robustness under realistic sharing conditions, we add 144,000 learned-compression variants at three quality levels. Each example includes platform, device, editing operation, compression metadata, bounding-box annotations, and edit descriptions in English, Bangla, and Persian/Farsi generated with five vision-language models (VLMs). We evaluate five separate benchmark VLMs on edit detection and edit-operation classification. By focusing on where and how consumer AI edits occur, rather than only whether an image is synthetic, ConsumerGenEdit provides a practical testbed for multilingual VLM grounding, image provenance, and robust interpretation of AI-edited media.
%U https://aclanthology.org/2026.emnlp-main.1661/
%P 37756-37789
Markdown (Informal)
[Edits in the Wild: A Multilingual Benchmark for Localizing On-Device Generative Image Edits on Consumer Platforms](https://aclanthology.org/2026.emnlp-main.1661/) (Soumik et al., EMNLP 2026)
ACL
- Mohd. Farhan Israk Soumik, Awal Ahmed Fime, Malithi Mithsara Wanniarachchi Kankanamge, Arun Babu Aryal, Hoda Joudi Haghighi, Bibek Koirala, Ahmed Imteaj, and Abdur R. Shahid. 2026. Edits in the Wild: A Multilingual Benchmark for Localizing On-Device Generative Image Edits on Consumer Platforms. In Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing, pages 37756–37789, Budapest, Hungary. Association for Computational Linguistics.