---
title: "Rethinking Composed Image Retrieval Evaluation: A"
description: "The authors reverse that process: they start from a textual edit and use image editing models to create the target image, so… 46 references."
authors: ["Tingyu Song", "Yanzhao Zhang", "Mingxin Li", "Zhuoning Guo", "Dingkun Long", "Pengjun Xie", "Siyue Zhang", "Yilun Zhao", "Shu Wu"]
year: 2026
venue: "Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)"
source: "https://aclanthology.org/2026.acl-long.2144/"
doi: "10.18653/v1/2026.acl-long.2144"
licence: "CC BY 4.0 (ACL Anthology)"
licence_url: "https://creativecommons.org/licenses/by/4.0/"
summary_licence: "Elamur Cosmic License"
canonical: "https://research.elamur.ai/free/en/papers/song-2026-rethinking-composed-image-retrieval-evaluation/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/song-2026-rethinking-composed-image-retrieval-evaluation/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2026](https://md.research.elamur.ai/free/en/papers/2026/page/47/) / Rethinking Composed Image Retrieval Evaluation: A Fine-Grained Benchmark from Image Editing

# Rethinking Composed Image Retrieval Evaluation: A Fine-Grained Benchmark from Image Editing

**Authors:** [Tingyu Song](https://md.research.elamur.ai/free/en/authors/tingyu-song/), [Yanzhao Zhang](https://md.research.elamur.ai/free/en/authors/yanzhao-zhang/), [Mingxin Li](https://md.research.elamur.ai/free/en/authors/mingxin-li/), Zhuoning Guo, [Dingkun Long](https://md.research.elamur.ai/free/en/authors/dingkun-long/), [Pengjun Xie](https://md.research.elamur.ai/free/en/authors/pengjun-xie/), [Siyue Zhang](https://md.research.elamur.ai/free/en/authors/siyue-zhang/), [Yilun Zhao](https://md.research.elamur.ai/free/en/authors/yilun-zhao/) and [Shu Wu](https://md.research.elamur.ai/free/en/authors/shu-wu/) · **Year:** 2026 · **Venue:** [Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)](https://md.research.elamur.ai/free/en/venues/acl/) ([ACL 2026](https://md.research.elamur.ai/free/en/venues/acl/2026/)) · **Publisher:** [Association for Computational Linguistics](https://md.research.elamur.ai/free/en/publishers/association-for-computational-linguistics/) · **Licence:** [CC BY 4.0 (ACL Anthology)](https://creativecommons.org/licenses/by/4.0/)

**Links:** [Publisher's page](https://aclanthology.org/2026.acl-long.2144/) · [PDF](https://aclanthology.org/2026.acl-long.2144.pdf) · [DOI 10.18653/v1/2026.acl-long.2144](https://doi.org/10.18653/v1/2026.acl-long.2144) · [OpenAlex](https://openalex.org/W7166828186)

**Citations:** 0 citations (OpenAlex, 2026-10-14)

## Summary
Existing Composed Image Retrieval (CIR) benchmarks such as CIRR, FashionIQ and CIRCO cover few modification types and have overlapping category labels. This happens because their queries are written after a target image has already been retrieved. The authors reverse that process: they start from a textual edit and use image editing models to create the target image, so the modification type is fixed in advance. They organize CIR needs into a taxonomy of five main categories and fifteen subcategories. Their pipeline filters seed images with an MLLM, generates and combines edit instructions, edits the images, and rewrites the instructions into natural queries, including negated and multi-condition ones. Two filtering stages and human validation follow. The result is EDIR, which pairs balanced queries with a large image gallery. Tests of 13 multimodal embedding models show uneven performance across subcategories. The paper also finds modality bias in earlier benchmarks. An in-domain training experiment separates weaknesses that more data can fix from those rooted in model design.

## Key points
- **EDIR size:** 5,000 queries and a gallery of 178,645 images, with categories evenly distributed. For comparison, CIRR has 4,148 queries and CIRCO has 1,000.
- **Model results:** Even strong models such as RzenEmbed and GME do not perform consistently across all 15 subcategories.
- **Modality bias:** MLLM-based models score higher on CIRCO with text-only input than with text and image together, so the reference image adds little there. EDIR requires using both modalities.
- **EDIR-MLLM:** This model is based on Qwen2.5-VL and trained on 225,000 synthesized triplets (15,000 per category), drawn from a pool of about 1.1M. It reaches 59.9% Recall@1, compared with a 36.9% average for other MLLM-based models.
- **Data-fixable vs. intrinsic weaknesses:** Color, material, texture and action improve sharply with targeted data. Count, spatial and viewpoint improve only modestly, which the authors attribute to intrinsic reasoning limits of current architectures.

*Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.*

## Institutions

- [Hong Kong University of Science and Technology](https://md.research.elamur.ai/free/en/universities/hong-kong-university-of-science-and-technology/) — University, Hong Kong, China · 170 works · 26,341 citations
- [Alibaba Group](https://md.research.elamur.ai/free/en/companies/alibaba-group/) — Company, Hangzhou, China · 201 works · 17,172 citations
- [Nanyang Technological University](https://md.research.elamur.ai/free/en/universities/nanyang-technological-university/) — University, Singapore, Singapore · 184 works · 27,484 citations
- [Yale University](https://md.research.elamur.ai/free/en/universities/yale-university/) — University, New Haven, United States · 101 works · 13,302 citations
- [University of Chinese Academy of Sciences](https://md.research.elamur.ai/free/en/universities/university-of-chinese-academy-of-sciences/) — University, Beijing, China · 188 works · 8,744 citations
- [The Hong Kong University of Science and Technology (Guangzhou)](https://md.research.elamur.ai/free/en/universities/the-hong-kong-university-of-science-and-technology-guangzhou/) — University, Guangzhou, China · 67 works · 1,765 citations
- [Alibaba Group (Cayman Islands)](https://md.research.elamur.ai/free/en/companies/alibaba-group-cayman-islands/) — Company, George Town, Cayman Islands · 36 works · 3,645 citations

📚 Shelves (3): [🖥️ Computer Science](https://md.research.elamur.ai/free/en/shelves/computer-science/) · 18,614 works · [👁️ Computer Vision and Pattern Recognition](https://md.research.elamur.ai/free/en/shelves/computer-science/computer-vision-and-pattern-recognition/) · 3,486 works · [🗄️ Information Systems](https://md.research.elamur.ai/free/en/shelves/computer-science/information-systems/) · 1,730 works

🧭 Topics (2): [📏 Evaluation & benchmarks](https://md.research.elamur.ai/free/en/topics/evaluation-benchmarks/) · 1,341 works · [🔍 Information retrieval & search](https://md.research.elamur.ai/free/en/topics/information-retrieval/) · 552 works

## Cite

Cite as: Tingyu Song, Yanzhao Zhang, Mingxin Li, Zhuoning Guo, Dingkun Long, Pengjun Xie, Siyue Zhang, Yilun Zhao and Shu Wu. 2026. Rethinking Composed Image Retrieval Evaluation: A Fine-Grained Benchmark from Image Editing. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers).

The paper itself, not this page. BibTeX from the library's record of the work.

```bibtex
@inproceedings{song2026rethinking,
  title = {{Rethinking Composed Image Retrieval Evaluation: A Fine-Grained Benchmark from Image Editing}},
  author = {Tingyu Song, Yanzhao Zhang, Mingxin Li, Zhuoning Guo, Dingkun Long, Pengjun Xie, Siyue Zhang, Yilun Zhao and Shu Wu},
  booktitle = {Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)},
  year = {2026},
  doi = {10.18653/v1/2026.acl-long.2144},
  url = {https://doi.org/10.18653/v1/2026.acl-long.2144}
}
```

- [← previous RetentiveKV: State-Space Memory for Uncertainty-Aware Multimodal KV Cache Eviction](https://md.research.elamur.ai/free/en/papers/liu-2026-retentivekv-state-space-memory-uncertainty/)
- [next → Rethinking Data Mixing from the Perspective of Large Language Models](https://md.research.elamur.ai/free/en/papers/xu-2026-rethinking-data-mixing-perspective-large/)
- [↑ back to papers of 2026](https://md.research.elamur.ai/free/en/papers/2026/page/47/)

## References

46 references, 11 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Kenan E Ak, Ashraf A Kassim, Joo Hwee Lim, and Jo Yew Tham. 2018. Learning attribute representations with localization for flexible fashion search. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 7708–7717. [find](https://scholar.google.com/scholar?q=Kenan+E+Ak%2C+Ashraf+A+Kassim%2C+Joo+Hwee+Lim%2C+and+Jo+Yew+Tham%2E+2018%2E+Learning+attribute+representations+with+localization+for+flexible+fashion+search%2E+In+Proceedings+of+the+IEEE+conference+on+computer+vi)
2. Shuai Bai, Keqin Chen, Xuejing Liu, Jialin Wang, Wenbin Ge, Sibo Song, Kai Dang, Peng Wang, Shijie Wang, Jun Tang, et al. 2025. Qwen2. 5-vl technical report. arXiv preprint arXiv:2502.13923. — [Qwen2.5-VL Technical Report](https://md.research.elamur.ai/free/en/papers/bai-2025-qwen2-5-vl-technical/)
3. Alberto Baldrati, Lorenzo Agnolucci, Marco Bertini, and Alberto Del Bimbo. 2023. Zero-shot composed image retrieval with textual inversion. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 15338–15347. [find](https://scholar.google.com/scholar?q=Alberto+Baldrati%2C+Lorenzo+Agnolucci%2C+Marco+Bertini%2C+and+Alberto+Del+Bimbo%2E+2023%2E+Zero%2Dshot+composed+image+retrieval+with+textual+inversion%2E+In+Proceedings+of+the+IEEE%2FCVF+International+Conference+on+C)
4. Tim Brooks, Aleksander Holynski, and Alexei A Efros. 2023. Instructpix2pix: Learning to follow image editing instructions. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 18392–18402. [find](https://scholar.google.com/scholar?q=Tim+Brooks%2C+Aleksander+Holynski%2C+and+Alexei+A+Efros%2E+2023%2E+Instructpix2pix%3A+Learning+to+follow+image+editing+instructions%2E+In+Proceedings+of+the+IEEE%2FCVF+conference+on+computer+vision+and+pattern+reco)
5. Haonan Chen, Liang Wang, Nan Yang, Yutao Zhu, Ziliang Zhao, Furu Wei, and Zhicheng Dou. 2025. mme5: Improving multimodal multilingual embeddings via high-quality synthetic data. arXiv preprint arXiv:2502.08468. [find](https://scholar.google.com/scholar?q=Haonan+Chen%2C+Liang+Wang%2C+Nan+Yang%2C+Yutao+Zhu%2C+Ziliang+Zhao%2C+Furu+Wei%2C+and+Zhicheng+Dou%2E+2025%2E+mme5%3A+Improving+multimodal+multilingual+embeddings+via+high%2Dquality+synthetic+data%2E+arXiv+preprint+arXiv%3A2)
6. Longye Du, Shuaiyu Deng, Ying Li, Jun Li, and Qi Tian. 2025. A survey on composed image retrieval. ACM Transactions on Multimedia Computing, Communications and Applications. [find](https://scholar.google.com/scholar?q=Longye+Du%2C+Shuaiyu+Deng%2C+Ying+Li%2C+Jun+Li%2C+and+Qi+Tian%2E+2025%2E+A+survey+on+composed+image+retrieval%2E+ACM+Transactions+on+Multimedia+Computing%2C+Communications+and+Applications%2E)
7. Chaoyou Fu, Peixian Chen, Yunhang Shen, Yulei Qin, Mengdan Zhang, Xu Lin, Jinrui Yang, Xiawu Zheng, Ke Li, Xing Sun, et al. 2025a. [Mme: A comprehensive evaluation benchmark for multimodal large language models](https://md.research.elamur.ai/free/en/papers/fu-2023-mme-comprehensive-evaluation-benchmark-multimodal/). In The Thirty-ninth Annual Conference on Neural Information Processing Systems Datasets a…
8. Chaoyou Fu, Yuhan Dai, Yongdong Luo, Lei Li, Shuhuai Ren, Renrui Zhang, Zihan Wang, Chenyu Zhou, Yunhang Shen, Mengdan Zhang, et al. 2025b. [Video-mme: The first-ever comprehensive evaluation benchmark of multi-modal llms in video analysis](https://md.research.elamur.ai/free/en/papers/fu-2024-video-mme-first-ever-comprehensive/). In Proceedings of the Computer Vision and Pattern Recognitio…
9. Geonmo Gu, Sanghyuk Chun, Wonjae Kim, , Yoohoon Kang, and Sangdoo Yun. 2024a. Language-only training of zero-shot composed image retrieval. In Conference on Computer Vision and Pattern Recognition (CVPR). [find](https://scholar.google.com/scholar?q=Geonmo+Gu%2C+Sanghyuk+Chun%2C+Wonjae+Kim%2C+%2C+Yoohoon+Kang%2C+and+Sangdoo+Yun%2E+2024a%2E+Language%2Donly+training+of+zero%2Dshot+composed+image+retrieval%2E+In+Conference+on+Computer+Vision+and+Pattern+Recognition+%28CV)
10. Geonmo Gu, Sanghyuk Chun, Wonjae Kim, HeeJae Jun, Yoohoon Kang, and Sangdoo Yun. 2024b. Compod-iff: Versatile composed image retrieval with latent diffusion. Transactions on Machine Learning Research. Expert Certification. [find](https://scholar.google.com/scholar?q=Geonmo+Gu%2C+Sanghyuk+Chun%2C+Wonjae+Kim%2C+HeeJae+Jun%2C+Yoohoon+Kang%2C+and+Sangdoo+Yun%2E+2024b%2E+Compod%2Diff%3A+Versatile+composed+image+retrieval+with+latent+diffusion%2E+Transactions+on+Machine+Learning+Research%2E)
11. Tiancheng Gu, Kaicheng Yang, Kaichen Zhang, Xiang An, Ziyong Feng, Yueyi Zhang, Weidong Cai, Jiankang Deng, and Lidong Bing. 2025. Unime-v2: Mllm-as-a-judge for universal multimodal embedding learning. arXiv preprint arXiv:2510.13515. — [arXiv:2510.13515](https://arxiv.org/abs/2510.13515) [find](https://scholar.google.com/scholar?q=Tiancheng+Gu%2C+Kaicheng+Yang%2C+Kaichen+Zhang%2C+Xiang+An%2C+Ziyong+Feng%2C+Yueyi+Zhang%2C+Weidong+Cai%2C+Jiankang+Deng%2C+and+Lidong+Bing%2E+2025%2E+Unime%2Dv2%3A+Mllm%2Das%2Da%2Djudge+for+universal+multimodal+embedding+learning)
12. Markus Hafner, Maria Katsantoni, Tino Köster, James Marks, Joyita Mukherjee, Dorothee Staiger, Jernej Ule, and Mihaela Zavolan. 2021. Clip and complementary methods. Nature Reviews Methods Primers, 1(1):20. [find](https://scholar.google.com/scholar?q=Markus+Hafner%2C+Maria+Katsantoni%2C+Tino+K%C3%B6ster%2C+James+Marks%2C+Joyita+Mukherjee%2C+Dorothee+Staiger%2C+Jernej+Ule%2C+and+Mihaela+Zavolan%2E+2021%2E+Clip+and+complementary+methods%2E+Nature+Reviews+Methods+Primers%2C+1%28)
13. Yuxin Hou, Eleonora Vig, Michael Donoser, and Loris Bazzani. 2021. Learning attribute-driven disentangled representations for interactive fashion retrieval. In Proceedings of the IEEE/CVF International conference on computer vision, pages 12147–12157. [find](https://scholar.google.com/scholar?q=Yuxin+Hou%2C+Eleonora+Vig%2C+Michael+Donoser%2C+and+Loris+Bazzani%2E+2021%2E+Learning+attribute%2Ddriven+disentangled+representations+for+interactive+fashion+retrieval%2E+In+Proceedings+of+the+IEEE%2FCVF+Internationa)
14. Chuong Huynh, Jinyu Yang, Ashish Tawari, Mubarak Shah, Son Tran, Raffay Hamid, Trishul Chilimbi, and Abhinav Shrivastava. 2025. Collm: A large language model for composed image retrieval. In Proceedings of the Computer Vision and Pattern Recognition Conference, pages 3994–4004. [find](https://scholar.google.com/scholar?q=Chuong+Huynh%2C+Jinyu+Yang%2C+Ashish+Tawari%2C+Mubarak+Shah%2C+Son+Tran%2C+Raffay+Hamid%2C+Trishul+Chilimbi%2C+and+Abhinav+Shrivastava%2E+2025%2E+Collm%3A+A+large+language+model+for+composed+image+retrieval%2E+In+Proceedin)
15. Weijian Jian, Yajun Zhang, Dawei Liang, Chunyu Xie, Yixiao He, Dawei Leng, and Yuhui Yin. 2025. Rzenembed: Towards comprehensive multimodal retrieval. arXiv preprint arXiv:2510.27350. — [arXiv:2510.27350](https://arxiv.org/abs/2510.27350) [find](https://scholar.google.com/scholar?q=Weijian+Jian%2C+Yajun+Zhang%2C+Dawei+Liang%2C+Chunyu+Xie%2C+Yixiao+He%2C+Dawei+Leng%2C+and+Yuhui+Yin%2E+2025%2E+Rzenembed%3A+Towards+comprehensive+multimodal+retrieval%2E+arXiv+preprint+arXiv%3A2510%2E27350%2E+%E2%80%94+%5BarXiv%3A2510%2E27)
16. Ting Jiang, Minghui Song, Zihan Zhang, Haizhen Huang, Weiwei Deng, Feng Sun, Qi Zhang, Deqing Wang, and Fuzhen Zhuang. 2024a. E5-v: Universal embeddings with multimodal large language models. arXiv preprint arXiv:2407.12580. — [arXiv:2407.12580](https://arxiv.org/abs/2407.12580) [find](https://scholar.google.com/scholar?q=Ting+Jiang%2C+Minghui+Song%2C+Zihan+Zhang%2C+Haizhen+Huang%2C+Weiwei+Deng%2C+Feng+Sun%2C+Qi+Zhang%2C+Deqing+Wang%2C+and+Fuzhen+Zhuang%2E+2024a%2E+E5%2Dv%3A+Universal+embeddings+with+multimodal+large+language+models%2E+arXiv+pr)
17. Ziyan Jiang, Rui Meng, Xinyi Yang, Semih Yavuz, Yingbo Zhou, and Wenhu Chen. 2024b. Vlm2vec: Training vision-language models for massive multimodal embedding tasks. arXiv preprint arXiv:2410.05160. — [arXiv:2410.05160](https://arxiv.org/abs/2410.05160) [find](https://scholar.google.com/scholar?q=Ziyan+Jiang%2C+Rui+Meng%2C+Xinyi+Yang%2C+Semih+Yavuz%2C+Yingbo+Zhou%2C+and+Wenhu+Chen%2E+2024b%2E+Vlm2vec%3A+Training+vision%2Dlanguage+models+for+massive+multimodal+embedding+tasks%2E+arXiv+preprint+arXiv%3A2410%2E05160%2E+%E2%80%94+)
18. Shyamgopal Karthik, Karsten Roth, Massimiliano Mancini, and Zeynep Akata. 2024. Vision-by-language for training-free compositional image retrieval. International Conference on Learning Representations (ICLR). [find](https://scholar.google.com/scholar?q=Shyamgopal+Karthik%2C+Karsten+Roth%2C+Massimiliano+Mancini%2C+and+Zeynep+Akata%2E+2024%2E+Vision%2Dby%2Dlanguage+for+training%2Dfree+compositional+image+retrieval%2E+International+Conference+on+Learning+Representations)
19. Junnan Li, Dongxu Li, Silvio Savarese, and Steven Hoi. 2023. [Blip-2: Bootstrapping language-image pre-training with frozen image encoders and large language models](https://md.research.elamur.ai/free/en/papers/li-2023-blip-2-bootstrapping-language-image/). In International conference on machine learning, pages 19730–19742. PMLR.
20. Junnan Li, Dongxu Li, Caiming Xiong, and Steven Hoi. 2022. [Blip: Bootstrapping language-image pre-training for unified vision-language understanding and generation](https://md.research.elamur.ai/free/en/papers/li-2022-blip-bootstrapping-language-image-pre/). In International conference on machine learning, pages 12888–12900. PMLR.
21. Sheng-Chieh Lin, Chankyu Lee, Mohammad Shoeybi, Jimmy Lin, Bryan Catanzaro, and Wei Ping. Mm-embed: Universal multimodal retrieval with multimodal llms. In The Thirteenth International Conference on Learning Representations. [find](https://scholar.google.com/scholar?q=Sheng%2DChieh+Lin%2C+Chankyu+Lee%2C+Mohammad+Shoeybi%2C+Jimmy+Lin%2C+Bryan+Catanzaro%2C+and+Wei+Ping%2E+Mm%2Dembed%3A+Universal+multimodal+retrieval+with+multimodal+llms%2E+In+The+Thirteenth+International+Conference+on+L)
22. Ruiqi Liu, Yi Han, Zhengbo Zhang, Li Yao, Zhiyuan Yan, Jialiang Shen, Zhijin Chen, Bo Sun, Lubin Weng, Jing Dong, Yan Wang, and Shu Wu. 2025. Beyond artifacts: Real-centric envelope modeling for reliable ai-generated image detection. ArXiv, abs/2512.20937. — [arXiv:2512.20937](https://arxiv.org/abs/2512.20937) [find](https://scholar.google.com/scholar?q=Ruiqi+Liu%2C+Yi+Han%2C+Zhengbo+Zhang%2C+Li+Yao%2C+Zhiyuan+Yan%2C+Jialiang+Shen%2C+Zhijin+Chen%2C+Bo+Sun%2C+Lubin+Weng%2C+Jing+Dong%2C+Yan+Wang%2C+and+Shu+Wu%2E+2025%2E+Beyond+artifacts%3A+Real%2Dcentric+envelope+modeling+for+relia)
23. Zheyuan Liu, Cristian Rodriguez-Opazo, Damien Teney, and Stephen Gould. 2021. Image retrieval on real-life images with pre-trained vision-and-language models. In Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), pages 2125– 2134. [find](https://scholar.google.com/scholar?q=Zheyuan+Liu%2C+Cristian+Rodriguez%2DOpazo%2C+Damien+Teney%2C+and+Stephen+Gould%2E+2021%2E+Image+retrieval+on+real%2Dlife+images+with+pre%2Dtrained+vision%2Dand%2Dlanguage+models%2E+In+Proceedings+of+the+IEEE%2FCVF+Internatio)
24. Rui Meng, Ziyan Jiang, Ye Liu, Mingyi Su, Xinyi Yang, Yuepeng Fu, Can Qin, Zeyuan Chen, Ran Xu, Caiming Xiong, et al. 2025. Vlm2vec-v2: Advancing multimodal embedding for videos, images, and visual documents. arXiv preprint arXiv:2507.04590. — [arXiv:2507.04590](https://arxiv.org/abs/2507.04590) [find](https://scholar.google.com/scholar?q=Rui+Meng%2C+Ziyan+Jiang%2C+Ye+Liu%2C+Mingyi+Su%2C+Xinyi+Yang%2C+Yuepeng+Fu%2C+Can+Qin%2C+Zeyuan+Chen%2C+Ran+Xu%2C+Caiming+Xiong%2C+et+al%2E+2025%2E+Vlm2vec%2Dv2%3A+Advancing+multimodal+embedding+for+videos%2C+images%2C+and+visual+do)
25. OpenSearch-AI. 2025. embedding-v1-7b. [find](https://scholar.google.com/scholar?q=OpenSearch%2DAI%2E+2025%2E+embedding%2Dv1%2D7b%2E)
26. Opensearch-ai/ops-mm-Bill Psomas, George Retsinas, Nikos Efthymiadis, Panagiotis Filntisis, Yannis Avrithis, Petros Maragos, Ondrej Chum, and Giorgos Tolias. 2025. Instance-level composed image retrieval. In The Thirty-ninth Annual Conference on Neural Information Processing Systems. [find](https://scholar.google.com/scholar?q=Opensearch%2Dai%2Fops%2Dmm%2DBill+Psomas%2C+George+Retsinas%2C+Nikos+Efthymiadis%2C+Panagiotis+Filntisis%2C+Yannis+Avrithis%2C+Petros+Maragos%2C+Ondrej+Chum%2C+and+Giorgos+Tolias%2E+2025%2E+Instance%2Dlevel+composed+image+retrie)
27. Kuniaki Saito, Kihyuk Sohn, Xiang Zhang, Chun-Liang Li, Chen-Yu Lee, Kate Saenko, and Tomas Pfister. 2023. Pic2word: Mapping pictures to words for zero-shot composed image retrieval. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 19305–19314. [find](https://scholar.google.com/scholar?q=Kuniaki+Saito%2C+Kihyuk+Sohn%2C+Xiang+Zhang%2C+Chun%2DLiang+Li%2C+Chen%2DYu+Lee%2C+Kate+Saenko%2C+and+Tomas+Pfister%2E+2023%2E+Pic2word%3A+Mapping+pictures+to+words+for+zero%2Dshot+composed+image+retrieval%2E+In+Proceedings+of)
28. Christoph Schuhmann, Richard Vencu, Romain Beaumont, Robert Kaczmarczyk, Clayton Mullis, Aarush Katta, Theo Coombes, Jenia Jitsev, and Aran Komatsuzaki. 2021. [Laion-400m: Open dataset of clip-filtered 400 million image-text pairs](https://md.research.elamur.ai/free/en/papers/schuhmann-2021-laion-400m-open-dataset-clip/). arXiv preprint arXiv:2111.02114.
29. Tingyu Song, Guo Gan, Mingsheng Shang, and Yilun Zhao. 2025a. Ifir: A comprehensive benchmark for evaluating instruction-following in expert-domain information retrieval. In North American Chapter of the Association for Computational Linguistics. [find](https://scholar.google.com/scholar?q=Tingyu+Song%2C+Guo+Gan%2C+Mingsheng+Shang%2C+and+Yilun+Zhao%2E+2025a%2E+Ifir%3A+A+comprehensive+benchmark+for+evaluating+instruction%2Dfollowing+in+expert%2Ddomain+information+retrieval%2E+In+North+American+Chapter+of+)
30. Xuemeng Song, Haoqiang Lin, Haokun Wen, Bohan Hou, Mingzhu Xu, and Liqiang Nie. 2025b. A comprehensive survey on composed image retrieval. ACM Transactions on Information Systems, 44(1):1– 54. [find](https://scholar.google.com/scholar?q=Xuemeng+Song%2C+Haoqiang+Lin%2C+Haokun+Wen%2C+Bohan+Hou%2C+Mingzhu+Xu%2C+and+Liqiang+Nie%2E+2025b%2E+A+comprehensive+survey+on+composed+image+retrieval%2E+ACM+Transactions+on+Information+Systems%2C+44%281%29%3A1%E2%80%93+54%2E)
31. Sagar Vaze, Nicolas Carion, and Ishan Misra. 2023. Genecis: A benchmark for general conditional image similarity. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 6862–6872. [find](https://scholar.google.com/scholar?q=Sagar+Vaze%2C+Nicolas+Carion%2C+and+Ishan+Misra%2E+2023%2E+Genecis%3A+A+benchmark+for+general+conditional+image+similarity%2E+In+Proceedings+of+the+IEEE%2FCVF+Conference+on+Computer+Vision+and+Pattern+Recognition%2C+)
32. Lucas Ventura, Antoine Yang, Cordelia Schmid, and Gül Varol. 2024. CoVR-2: Automatic data construction for composed video retrieval. IEEE TPAMI. [find](https://scholar.google.com/scholar?q=Lucas+Ventura%2C+Antoine+Yang%2C+Cordelia+Schmid%2C+and+G%C3%BCl+Varol%2E+2024%2E+CoVR%2D2%3A+Automatic+data+construction+for+composed+video+retrieval%2E+IEEE+TPAMI%2E)
33. Yongquan Wan, Guobing Zou, and Bofeng Zhang. 2025. Composed image retrieval: a survey on recent research and development. Applied Intelligence, 55(6):482. [find](https://scholar.google.com/scholar?q=Yongquan+Wan%2C+Guobing+Zou%2C+and+Bofeng+Zhang%2E+2025%2E+Composed+image+retrieval%3A+a+survey+on+recent+research+and+development%2E+Applied+Intelligence%2C+55%286%29%3A482%2E)
34. Peng Wang, Shuai Bai, Sinan Tan, Shijie Wang, Zhihao Fan, Jinze Bai, Keqin Chen, Xuejing Liu, Jialin Wang, Wenbin Ge, et al. 2024. [Qwen2-vl: Enhancing vision-language model’s perception of the world at any resolution](https://md.research.elamur.ai/free/en/papers/wang-2024-qwen2-vl-enhancing-vision/). arXiv preprint arXiv:2409.12191.
35. Chenfei Wu, Jiahao Li, Jingren Zhou, Junyang Lin, Kaiyuan Gao, Kun Yan, Sheng ming Yin, Shuai Bai, Xiao Xu, Yilei Chen, Yuxiang Chen, Zecheng Tang, Zekai Zhang, Zhengyi Wang, An Yang, Bowen Yu, Chen Cheng, Dayiheng Liu, Deqing Li, Hang Zhang, Hao Meng, Hu Wei, Jingyuan Ni, Kai Chen, Kuan Cao, Liang… — [Qwen-Image Technical Report](https://md.research.elamur.ai/free/en/papers/wu-2025-qwen-image-technical-report/)
36. Hui Wu, Yupeng Gao, Xiaoxiao Guo, Ziad Al-Halah, Steven Rennie, Kristen Grauman, and Rogerio Feris. 2021. Fashion iq: A new dataset towards retrieving images by natural language feedback. In Proceedings of the IEEE/CVF Conference on computer vision and pattern recognition, pages 11307–11317. [find](https://scholar.google.com/scholar?q=Hui+Wu%2C+Yupeng+Gao%2C+Xiaoxiao+Guo%2C+Ziad+Al%2DHalah%2C+Steven+Rennie%2C+Kristen+Grauman%2C+and+Rogerio+Feris%2E+2021%2E+Fashion+iq%3A+A+new+dataset+towards+retrieving+images+by+natural+language+feedback%2E+In+Proceedin)
37. Ren-Di Wu, Yu-Yen Lin, and Huei-Fang Yang. 2024. Training-free zero-shot composed image retrieval via weighted modality fusion and similarity. In International Conference on Technologies and Applications of Artificial Intelligence, pages 77–90. Springer. [find](https://scholar.google.com/scholar?q=Ren%2DDi+Wu%2C+Yu%2DYen+Lin%2C+and+Huei%2DFang+Yang%2E+2024%2E+Training%2Dfree+zero%2Dshot+composed+image+retrieval+via+weighted+modality+fusion+and+similarity%2E+In+International+Conference+on+Technologies+and+Applicati)
38. An Yang, Anfeng Li, Baosong Yang, Beichen Zhang, Binyuan Hui, Bo Zheng, Bowen Yu, Chang Gao, Chengen Huang, Chenxu Lv, et al. 2025. [Qwen3 technical report](https://md.research.elamur.ai/free/en/papers/yang-2025-qwen3-technical-report/). arXiv preprint arXiv:2505.09388.
39. Xin Yang, Xuemeng Song, Xianjing Han, Haokun Wen, Jie Nie, and Liqiang Nie. 2020. Generative attribute manipulation scheme for flexible fashion search. In Proceedings of the 43rd international acm sigir conference on research and development in information retrieval, pages 941–950. [find](https://scholar.google.com/scholar?q=Xin+Yang%2C+Xuemeng+Song%2C+Xianjing+Han%2C+Haokun+Wen%2C+Jie+Nie%2C+and+Liqiang+Nie%2E+2020%2E+Generative+attribute+manipulation+scheme+for+flexible+fashion+search%2E+In+Proceedings+of+the+43rd+international+acm+sig)
40. Zhenyu Yang, Dizhan Xue, Shengsheng Qian, Weiming Dong, and Changsheng Xu. 2024. Ldre: Llm-based divergent reasoning and ensemble for zero-shot composed image retrieval. In Proceedings of the 47th International ACM SIGIR conference on research and development in information retrieval, pages 80– 90. [find](https://scholar.google.com/scholar?q=Zhenyu+Yang%2C+Dizhan+Xue%2C+Shengsheng+Qian%2C+Weiming+Dong%2C+and+Changsheng+Xu%2E+2024%2E+Ldre%3A+Llm%2Dbased+divergent+reasoning+and+ensemble+for+zero%2Dshot+composed+image+retrieval%2E+In+Proceedings+of+the+47th+Int)
41. Kai Zhang, Yi Luan, Hexiang Hu, Kenton Lee, Siyuan Qiao, Wenhu Chen, Yu Su, and Ming-Wei Chang. Magiclens: Self-supervised image retrieval with open-ended instructions. In Forty-first International Conference on Machine Learning. [find](https://scholar.google.com/scholar?q=Kai+Zhang%2C+Yi+Luan%2C+Hexiang+Hu%2C+Kenton+Lee%2C+Siyuan+Qiao%2C+Wenhu+Chen%2C+Yu+Su%2C+and+Ming%2DWei+Chang%2E+Magiclens%3A+Self%2Dsupervised+image+retrieval+with+open%2Dended+instructions%2E+In+Forty%2Dfirst+International+Co)
42. Siyue Zhang, Yuan Gao, Xiao Zhou, Yilun Zhao, Tingyu Song, Arman Cohan, Anh Tuan Luu, and Chen Zhao. 2025. Mrmr: A realistic and expert-level multidisciplinary benchmark for reasoning-intensive multimodal retrieval. ArXiv, abs/2510.09510. — [arXiv:2510.09510](https://arxiv.org/abs/2510.09510) [find](https://scholar.google.com/scholar?q=Siyue+Zhang%2C+Yuan+Gao%2C+Xiao+Zhou%2C+Yilun+Zhao%2C+Tingyu+Song%2C+Arman+Cohan%2C+Anh+Tuan+Luu%2C+and+Chen+Zhao%2E+2025%2E+Mrmr%3A+A+realistic+and+expert%2Dlevel+multidisciplinary+benchmark+for+reasoning%2Dintensive+multim)
43. Xin Zhang, Yanzhao Zhang, Wen Xie, Mingxin Li, Ziqi Dai, Dingkun Long, Pengjun Xie, Meishan Zhang, Wenjie Li, and Min Zhang. 2024. Gme: Improving universal multimodal retrieval by multimodal llms. arXiv preprint arXiv:2412.16855. — [arXiv:2412.16855](https://arxiv.org/abs/2412.16855) [find](https://scholar.google.com/scholar?q=Xin+Zhang%2C+Yanzhao+Zhang%2C+Wen+Xie%2C+Mingxin+Li%2C+Ziqi+Dai%2C+Dingkun+Long%2C+Pengjun+Xie%2C+Meishan+Zhang%2C+Wenjie+Li%2C+and+Min+Zhang%2E+2024%2E+Gme%3A+Improving+universal+multimodal+retrieval+by+multimodal+llms%2E+arX)
44. Junjie Zhou, Zheng Liu, Shitao Xiao, Bo Zhao, and Yongping Xiong. 2024. Vista: Visualized text embedding for universal multi-modal retrieval. arXiv preprint arXiv:2406.04292. [find](https://scholar.google.com/scholar?q=Junjie+Zhou%2C+Zheng+Liu%2C+Shitao+Xiao%2C+Bo+Zhao%2C+and+Yongping+Xiong%2E+2024%2E+Vista%3A+Visualized+text+embedding+for+universal+multi%2Dmodal+retrieval%2E+arXiv+preprint+arXiv%3A2406%2E04292%2E)
45. Junjie Zhou, Yongping Xiong, Zheng Liu, Ze Liu, Shitao Xiao, Yueze Wang, Bo Zhao, Chen Jason Zhang, and Defu Lian. 2025. [Megapairs: Massive data synthesis for universal multimodal retrieval](https://md.research.elamur.ai/free/en/papers/zhou-2025-megapairs-massive-data-synthesis-universal/). In Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Pa…
46. Jinguo Zhu, Weiyun Wang, Zhe Chen, Zhaoyang Liu, Shenglong Ye, Lixin Gu, Hao Tian, Yuchen Duan, Weijie Su, Jie Shao, et al. 2025. [Internvl3: Exploring advanced training and test-time recipes for open-source multimodal models](https://md.research.elamur.ai/free/en/papers/zhu-2025-internvl3-exploring-advanced-training/). arXiv preprint arXiv:2504.10479. *Figure 5: Example of Error Type: Failure…

*Updated 2 November 2026*