pith. machine review for the scientific record. sign in

arxiv: 2604.09249 · v2 · submitted 2026-04-10 · 💻 cs.CV · cs.IR

Recognition: no theorem link

FashionStylist: An Expert Knowledge-enhanced Multimodal Dataset for Fashion Understanding

Kaidong Feng , Zhuoxuan Huang , Huizhong Guo , Yuting Jin , Xinyu Chen , Yue Liang , Yifei Gai , Li Zhou
show 2 more authors
Yunshan Ma Zhu Sun
Authors on Pith no claims yet

Pith reviewed 2026-05-10 18:05 UTC · model grok-4.3

classification 💻 cs.CV cs.IR
keywords fashion datasetexpert annotationsoutfit understandingmultimodal learningitem groundingoutfit completionoutfit evaluationvision-language models
0
0 comments X

The pith

FashionStylist supplies expert-annotated data to benchmark and train models on three outfit-level fashion tasks.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper introduces a multimodal dataset built with professional fashion annotations at both item and outfit scales to overcome the limits of earlier collections that focused only on isolated attributes or loose text labels. It defines three tasks that require recovering hidden items from layered ensembles, assembling compatible new pieces, and scoring an outfit for style, season, occasion, and internal fit. If the annotations prove superior and the tasks prove representative, models can move from superficial matching to coherent, expert-style reasoning about clothing. The work therefore positions the dataset as both a common testbed and a direct training source for multimodal systems handling fashion.

Core claim

FashionStylist is assembled via a dedicated expert annotation pipeline that yields grounded labels for items inside complex outfits and for whole-outfit properties. The three supported tasks are outfit-to-item grounding that recovers specific garments amid layering and accessories, compatibility-aware outfit completion that goes beyond simple co-occurrence, and outfit evaluation that assesses style, season, occasion, and overall coherence. Experiments indicate the resource functions simultaneously as a unified benchmark across these tasks and as training material that raises performance in grounding, completion, and semantic evaluation for multimodal large language model fashion systems.

What carries the argument

The expert annotation pipeline that supplies item-level and outfit-level labels for the three tasks of grounding, completion, and evaluation.

If this is right

  • Models gain improved recovery of individual items from outfits that include layering and accessories.
  • Outfit completion becomes driven by compatibility rules rather than frequency of past pairings.
  • Outfit evaluation incorporates expert judgment on style, season, occasion, and internal coherence.
  • A single annotated collection can replace multiple task-specific datasets for both testing and fine-tuning.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same expert-pipeline approach could be replicated for other visual domains that combine appearance with cultural rules, such as interior styling or product photography.
  • Wider adoption might shift fashion AI away from scraping social media tags toward curated professional knowledge.
  • Direct ablation studies that swap expert labels for weak labels on the same images would isolate the contribution of annotation quality.
  • Integration with generative models could test whether the dataset improves synthesized outfits that satisfy the same coherence criteria.

Load-bearing premise

Expert annotations produced by the dedicated pipeline are reliably better than weak textual supervision and the three tasks together capture holistic fashion understanding.

What would settle it

Training identical multimodal models on prior fashion collections versus FashionStylist and observing no measurable gain in grounding accuracy, completion quality, or agreement with expert outfit scores.

Figures

Figures reproduced from arXiv: 2604.09249 by Huizhong Guo, Kaidong Feng, Li Zhou, Xinyu Chen, Yifei Gai, Yue Liang, Yunshan Ma, Yuting Jin, Zhuoxuan Huang, Zhu Sun.

Figure 1
Figure 1. Figure 1: Overview of our proposed FashionStylist, where the [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: (Left) Number of unique attribute values in Fash￾ionStylist across item- and outfit-level annotations. (Right) Normalized co-occurrence frequency between the top-5 most common colors and outfit styles. To ensure consistency, the annotation process follows an iterative workflow in which annotation guidelines are continuously refined through cross-review and consistency verification, together with initial la… view at source ↗
Figure 3
Figure 3. Figure 3: Performance comparison across two representative [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
read the original abstract

Fashion understanding requires both visual perception and expert-level reasoning about style, occasion, compatibility, and outfit rationale. However, existing fashion datasets remain fragmented and task-specific, often focusing on item attributes, outfit co-occurrence, or weak textual supervision, and thus provide limited support for holistic outfit understanding. In this paper, we introduce FashionStylist, an expert-annotated benchmark for holistic and expert-level fashion understanding. Constructed through a dedicated fashion-expert annotation pipeline, FashionStylist provides professionally grounded annotations at both the item and outfit levels. It supports three representative tasks: outfit-to-item grounding, outfit completion, and outfit evaluation. These tasks cover realistic item recovery from complex outfits with layering and accessories, compatibility-aware composition beyond co-occurrence matching, and expert-level assessment of style, season, occasion, and overall coherence. Experimental results show that FashionStylist serves not only as a unified benchmark for multiple fashion tasks, but also as an effective training resource for improving grounding, completion, and outfit-level semantic evaluation in MLLM-based fashion systems.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 1 minor

Summary. The paper introduces FashionStylist, an expert-annotated multimodal dataset for holistic fashion understanding constructed via a dedicated fashion-expert annotation pipeline. It supports three tasks—outfit-to-item grounding, outfit completion, and outfit evaluation—and claims that experimental results establish it as both a unified benchmark for multiple fashion tasks and an effective training resource for improving grounding, completion, and outfit-level semantic evaluation in MLLM-based systems.

Significance. If the experimental claims are substantiated, the dataset would address fragmentation in existing fashion resources by providing professionally grounded item- and outfit-level annotations, potentially enabling more robust multimodal models for style, compatibility, and coherence reasoning. The emphasis on expert-level tasks beyond weak supervision or co-occurrence matching could serve as a valuable benchmark and training resource in the fashion CV and MLLM communities.

major comments (2)
  1. [Abstract] Abstract: The central claim that 'experimental results show that FashionStylist serves ... as an effective training resource for improving grounding, completion, and outfit-level semantic evaluation' is unsupported, as the manuscript provides no quantitative metrics, baselines, ablation studies, data splits, or performance comparisons on the three tasks. This directly undermines verification of the training-resource assertion.
  2. [Experimental evaluation] Experimental evaluation section: No inter-annotator agreement scores, comparisons against prior weak-supervision datasets, or results tables are presented to demonstrate that the expert annotation pipeline yields reliably superior labels or measurable gains in MLLM performance, leaving the superiority assumption untested.
minor comments (1)
  1. The description of the three tasks would benefit from additional concrete examples or illustrative figures to clarify distinctions from prior task-specific datasets.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive feedback on our manuscript introducing FashionStylist. The comments highlight important gaps in substantiating our claims about the dataset's utility as both a benchmark and training resource. We address each point below and will revise the manuscript to incorporate the requested evidence.

read point-by-point responses

  1. Referee: [Abstract] Abstract: The central claim that 'experimental results show that FashionStylist serves ... as an effective training resource for improving grounding, completion, and outfit-level semantic evaluation' is unsupported, as the manuscript provides no quantitative metrics, baselines, ablation studies, data splits, or performance comparisons on the three tasks. This directly undermines verification of the training-resource assertion.

    Authors: We acknowledge that the current manuscript version presents the dataset construction, task definitions, and high-level experimental claims without the detailed quantitative support referenced in the abstract. To rectify this, we will add a comprehensive experimental section that includes quantitative metrics (e.g., accuracy, IoU for grounding; compatibility scores for completion; coherence ratings for evaluation), baseline MLLM comparisons, ablation studies on training with FashionStylist versus other data, data splits, and performance gains demonstrating its value as a training resource. revision: yes

  2. Referee: [Experimental evaluation] Experimental evaluation section: No inter-annotator agreement scores, comparisons against prior weak-supervision datasets, or results tables are presented to demonstrate that the expert annotation pipeline yields reliably superior labels or measurable gains in MLLM performance, leaving the superiority assumption untested.

    Authors: We agree that explicit validation of the expert annotation pipeline's quality is necessary to support claims of superiority over weak-supervision approaches. In the revised manuscript, we will include inter-annotator agreement scores (e.g., Cohen's kappa or percentage agreement) for key annotation aspects, direct comparisons against prior fashion datasets relying on weak supervision or co-occurrence, and results tables quantifying MLLM performance improvements attributable to FashionStylist's expert labels. revision: yes

Circularity Check

0 steps flagged

No circularity: dataset introduction with no derivations or self-referential predictions

full rationale

The paper introduces FashionStylist as a new expert-annotated multimodal dataset supporting three tasks (outfit-to-item grounding, outfit completion, outfit evaluation). No equations, fitted parameters, or first-principles derivations are present in the provided text. Claims rest on dataset construction via an expert pipeline and unspecified experimental results, without any reduction of outputs to inputs by definition, self-citation chains, or renaming of known results. This matches the default expectation for a dataset paper: self-contained contribution without circular steps.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The central claim rests on the premise that expert annotations add value beyond existing supervision; no free parameters or invented entities are introduced.

axioms (1)
  • domain assumption Expert fashion knowledge can be captured through a dedicated annotation pipeline to produce higher-quality labels than weak textual supervision for multimodal models.
    Invoked in the description of the annotation process and the claimed benefits for training and benchmarking.

pith-pipeline@v0.9.0 · 5510 in / 1150 out tokens · 77793 ms · 2026-05-10T18:05:10.321250+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

55 extracted references · 12 canonical work pages · 7 internal anchors

  1. [1]

    Shuai Bai, Yuxuan Cai, Ruizhe Chen, Keqin Chen, Xionghui Chen, Zesen Cheng, Lianghao Deng, Wei Ding, et al . 2025. Qwen3-VL technical report. arXiv:2511.21631 [cs.CV] https://arxiv.org/abs/2511.21631

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  2. [2]

    Shuai Bai, Keqin Chen, Xuejing Liu, Jialin Wang, Wenbin Ge, Sibo Song, Kai Dang, Peng Wang, et al. 2025. Qwen2.5-VL technical report. arXiv:2502.13923 [cs.CV] https://arxiv.org/abs/2502.13923

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  3. [3]

    Mikołaj Bińkowski, Danica J Sutherland, Michael Arbel, and Arthur Gretton

  4. [4]

    Demystifying mmd gans.arXiv preprint arXiv:1801.01401(2018)

    work page internal anchor Pith review arXiv 2018

  5. [5]

    Wen Chen, Pipei Huang, Jiaming Xu, Xin Guo, Cheng Guo, Fei Sun, Chao Li, Andreas Pfadler, Huan Zhao, and Binqiang Zhao. 2019. POG: personalized outfit generation for fashion recommendation at Alibaba iFashion. InProceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining. 2662–2670

    2019

  6. [6]

    Wen-Huang Cheng, Sijie Song, Chieh-Yun Chen, Shintami Chusnul Hidayati, and Jiaying Liu. 2021. Fashion meets computer vision: A survey.Comput. Surveys 54, 4 (2021), 1–41

    2021

  7. [7]

    Patrick John Chia, Giuseppe Attanasio, Federico Bianchi, Silvia Terragni, Ana Rita Magalhaes, Diogo Goncalves, Ciro Greco, and Jacopo Tagliabue. 2022. Contrastive language and vision learning of general fashion concepts.Scientific Reports12, 1 (2022), 18958

    2022

  8. [8]

    Seunghwan Choi, Sunghyun Park, Minsoo Lee, and Jaegul Choo. 2021. Viton- hd: High-resolution virtual try-on via misalignment-aware normalization. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 14131–14140

    2021

  9. [9]

    Jacob Cohen. 1960. A coefficient of agreement for nominal scales.Educational and psychological measurement20, 1 (1960), 37–46

    1960

  10. [10]

    Gayatri Deshmukh, Somsubhra De, Chirag Sehgal, Jishu Sen Gupta, and Sparsh Mittal. 2024. Dressing the imagination: A Dataset for AI-powered translation of text into fashion outfits and a novel NeRA adapter for enhanced feature adaptation.arXiv preprint arXiv:2411.13901(2024)

    work page arXiv 2024

  11. [11]

    Tim Dettmers, Artidoro Pagnoni, Ari Holtzman, and Luke Zettlemoyer. 2023. Qlora: Efficient finetuning of quantized llms.Advances in Neural Information Processing Systems(2023), 10088–10115

    2023

  12. [12]

    Yujuan Ding, Zhihui Lai, PY Mok, and Tat-Seng Chua. 2023. Computational technologies for fashion recommendation: A survey.Comput. Surveys56, 5 (2023), 1–45

    2023

  13. [13]

    Yuying Ge, Ruimao Zhang, Xiaogang Wang, Xiaoou Tang, and Ping Luo. 2019. Deepfashion2: A versatile benchmark for detection, pose estimation, segmen- tation and re-identification of clothing images. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 5337–5345

    2019

  14. [14]

    Weili Guan, Fangkai Jiao, Xuemeng Song, Haokun Wen, Chung-Hsing Yeh, and Xiaojun Chang. 2022. Personalized fashion compatibility modeling via metapath- guided heterogeneous graph learning. InProceedings of the 45th international ACM SIGIR Conference on Research and Development in Information Retrieval. 482–491

    2022

  15. [15]

    Martin Heusel, Hubert Ramsauer, Thomas Unterthiner, Bernhard Nessler, and Sepp Hochreiter. 2017. GANs trained by a two time-scale update rule converge to a local nash equilibrium.Advances in Neural Information Processing Systems (2017)

    2017

  16. [16]

    Edward J Hu, Yelong Shen, Phillip Wallis, Zeyuan Allen-Zhu, Yuanzhi Li, Shean Wang, Lu Wang, and Weizhu Chen. 2022. LoRA: Low-rank adaptation of large language models. InProceedings of the International Conference on Learning Representations (ICLR)

    2022

  17. [17]

    Youngseung Jeon, Seungwan Jin, Patrick C Shih, and Kyungsik Han. 2021. Fash- ionQ: An ai-driven creativity support tool for facilitating ideation in fashion design. InProceedings of the 2021 CHI Conference on Human Factors in Computing Systems. 1–18

    2021

  18. [18]

    Menglin Jia, Mengyun Shi, Mikhail Sirotenko, Yin Cui, Claire Cardie, Bharath Hariharan, Hartwig Adam, and Serge Belongie. 2020. Fashionpedia: Ontology, segmentation, and an attribute localization dataset. InEuropean Conference on Computer Vision. Springer, 316–332

    2020

  19. [19]

    Peng Jin, Yilin Wen, Mingzhe Yu, Yunshan Ma, Rong Zheng, Jin-tu Fan, and Chong Wah NGO. 2025. GenWardrobe: A fully generative system for travel fashion wardrobe construction. InProceedings of the 33rd ACM International Conference on Multimedia. 13540–13542

    2025

  20. [20]

    Onur Keleş, M Akın Yılmaz, A Murat Tekalp, Cansu Korkmaz, and Zafer Dogan

  21. [21]

    On the computation of PSNR for a set of images or video.arXiv preprint arXiv:2104.14868(2021)

    work page arXiv 2021

  22. [22]

    Black Forest Labs, Stephen Batifol, Andreas Blattmann, Frederic Boesel, Saksham Consul, Cyril Diagne, Tim Dockhorn, Jack English, Zion English, Patrick Esser, et al. 2025. FLUX. 1 Kontext: Flow matching for in-context image generation and editing in latent space.arXiv preprint arXiv:2506.15742(2025)

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  23. [23]

    J Richard Landis and Gary G Koch. 1977. The measurement of observer agreement for categorical data.biometrics(1977), 159–174

    1977

  24. [24]

    Lizi Liao, You Zhou, Yunshan Ma, Richang Hong, and Tat-seng Chua. 2018. Knowledge-aware multimodal fashion chatbot. InProceedings of the 26th ACM International Conference on Multimedia. 1265–1266

    2018

  25. [25]

    Yen-Liang Lin, Son Tran, and Larry S Davis. 2020. Fashion outfit complementary item retrieval. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 3311–3319

    2020

  26. [26]

    Xiaohao Liu, Jie Wu, Zhulin Tao, Yunshan Ma, Yinwei Wei, and Tat-seng Chua

  27. [27]

    In Proceedings of the 31st ACM SIGKDD Conference on Knowledge Discovery and Data Mining

    Fine-tuning multimodal large language models for product bundling. In Proceedings of the 31st ACM SIGKDD Conference on Knowledge Discovery and Data Mining. 848–858

  28. [28]

    Ziwei Liu, Ping Luo, Shi Qiu, Xiaogang Wang, and Xiaoou Tang. 2016. Deep- Fashion: Powering robust clothes recognition and retrieval with rich annotations. In2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). 1096–1104. https://doi.org/10.1109/CVPR.2016.124

    work page doi:10.1109/cvpr.2016.124 2016

  29. [29]

    Zhi Lu, Yang Hu, Yunchao Jiang, Yan Chen, and Bing Zeng. 2019. Learning binary code for personalized fashion recommendation. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 10562–10570

    2019

  30. [30]

    Yunshan Ma, Yujuan Ding, Xun Yang, Lizi Liao, Wai Keung Wong, and Tat-Seng Chua. 2020. Knowledge enhanced neural fashion trend forecasting. InProceedings of the 2020 International Conference on Multimedia Retrieval. 82–90

    2020

  31. [31]

    Yunshan Ma, Yingzhi He, Wenjun Zhong, Xiang Wang, Roger Zimmermann, and Tat-Seng Chua. 2024. Cirp: Cross-item relational pre-training for multimodal product bundling. InProceedings of the 32nd ACM International Conference on Multimedia. 9641–9649

    2024

  32. [32]

    Yunshan Ma, Xiaohao Liu, Yinwei Wei, Zhulin Tao, Xiang Wang, and Tat-Seng Chua. 2024. Leveraging multimodal features and item-level user feedback for bundle construction. InProceedings of the 17th ACM International Conference on Web Search and Data Mining. 510–519

    2024

  33. [33]

    Davide Morelli, Matteo Fincato, Marcella Cornia, Federico Landi, Fabio Cesari, and Rita Cucchiara. 2022. Dress code: High-resolution multi-category virtual try-on. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 2231–2235

    2022

  34. [34]

    Jianmo Ni, Jiacheng Li, and Julian McAuley. 2019. Justifying recommendations using distantly-labeled reviews and fine-grained aspects. InProceedings of the Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing. 188–197

    2019

  35. [35]

    Negar Rostamzadeh, Seyedarian Hosseini, Thomas Boquet, Wojciech Stokowiec, Ying Zhang, Christian Jauvin, and Chris Pal. 2018. Fashion-gen: The generative fashion dataset and challenge.arXiv preprint arXiv:1806.08317(2018)

    work page arXiv 2018

  36. [36]

    Burak Satar, Zhixin Ma, Patrick Amadeus Irawan, Wilfried Ariel Mulyawan, Jing Jiang, Ee-Peng Lim, and Chong-Wah Ngo. 2025. Seeing culture: A benchmark for visual reasoning and grounding. InProceedings of the 2025 Conference on Empirical Methods in Natural Language Processing. 22238–22254

    2025

  37. [37]

    Wenda Shi, Waikeung Wong, and Xingxing Zou. 2025. Generative AI in fashion: Overview.ACM Transactions on Intelligent Systems and Technology16, 4 (2025), 1–73

    2025

  38. [38]

    Xuemeng Song, Xianjing Han, Yunkai Li, Jingyuan Chen, Xin-Shun Xu, and Liqiang Nie. 2019. GP-BPR: Personalized compatibility modeling for clothing matching. InProceedings of the 27th ACM international Conference on Multimedia. 320–328

    2019

  39. [39]

    Tianyu Su, Xuemeng Song, Na Zheng, Weili Guan, Yan Li, and Liqiang Nie

  40. [40]

    In Proceedings of the 29th ACM International Conference on Multimedia

    Complementary factorization towards outfit compatibility modeling. In Proceedings of the 29th ACM International Conference on Multimedia. 4073–4081

  41. [41]

    Zhu Sun, Kaidong Feng, Jie Yang, Xinghua Qu, Hui Fang, Yew-Soon Ong, and Wenyuan Liu. 2024. Adaptive in-context learning with large language models for bundle generation. InProceedings of the 47th International ACM SIGIR Conference on Research and Development in Information Retrieval. 966–976

    2024

  42. [42]

    Gemma Team. 2025. Gemma 3 technical report. arXiv:2503.19786 [cs.CL] https://arxiv.org/abs/2503.19786

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  43. [43]

    Gemini Team, Rohan Anil, Sebastian Borgeaud, Jean-Baptiste Alayrac, Jiahui Yu, Radu Soricut, Johan Schalkwyk, Andrew M Dai, Anja Hauth, Katie Millican, et al

  44. [44]

    Gemini: A Family of Highly Capable Multimodal Models

    Gemini: A family of highly capable multimodal models. arXiv 2023.arXiv preprint arXiv:2312.11805(2024)

    work page internal anchor Pith review Pith/arXiv arXiv 2023

  45. [45]

    Meituan LongCat Team, Hanghang Ma, Haoxian Tan, Jiale Huang, Junqiang Wu, Jun-Yan He, Lishuai Gao, Songlin Xiao, Xiaoming Wei, Xiaoqi Ma, et al. 2025. Longcat-image technical report.arXiv preprint arXiv:2512.07584(2025)

    work page arXiv 2025

  46. [46]

    Xin Wang, Bo Wu, and Yueqi Zhong. 2019. Outfit compatibility prediction and diagnosis with multi-layered comparison network. InProceedings of the 27th ACM International Conference on Multimedia. 329–337

    2019

  47. [47]

    Zhou Wang, Alan C Bovik, Hamid R Sheikh, and Eero P Simoncelli. 2004. Image quality assessment: From error visibility to structural similarity.IEEE Transactions on Image Processing13, 4 (2004), 600–612

    2004

  48. [48]

    Chenfei Wu, Jiahao Li, Jingren Zhou, Junyang Lin, Kaiyuan Gao, Kun Yan, Sheng- ming Yin, Shuai Bai, Xiao Xu, Yilei Chen, et al . 2025. Qwen-image technical report.arXiv preprint arXiv:2508.02324(2025)

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  49. [49]

    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. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 11307–11317. xxx ’17, November 11–24, 2006, xxx, xxxx Feng et al

    2021

  50. [50]

    Yiyan Xu, Wenjie Wang, Fuli Feng, Yunshan Ma, Jizhi Zhang, and Xiangnan He

  51. [51]

    InProceedings of the 47th International ACM SIGIR Conference on Research and Development in Information Retrieval

    Diffusion models for generative outfit recommendation. InProceedings of the 47th International ACM SIGIR Conference on Research and Development in Information Retrieval. 1350–1359

  52. [52]

    Xuewen Yang, Dongliang Xie, Xin Wang, Jiangbo Yuan, Wanying Ding, and Pengyun Yan. 2020. Learning tuple compatibility for conditional outfit recom- mendation. InProceedings of the 28th ACM International Conference on Multimedia. 2636–2644

    2020

  53. [53]

    Xiangyu Zhao, Yuehan Zhang, Wenlong Zhang, and Xiao-Ming Wu. 2024. Uni- fashion: A unified vision-language model for multimodal fashion retrieval and generation. InProceedings of the 2024 Conference on Empirical Methods in Natural Language Processing. 1490–1507

    2024

  54. [54]

    Haitian Zheng, Kefei Wu, Jong-Hwi Park, Wei Zhu, and Jiebo Luo. 2021. Person- alized fashion recommendation from personal social media data: An item-to-set metric learning approach. In2021 IEEE International Conference on Big Data. IEEE, 5014–5023

    2021

  55. [55]

    Dongliang Zhou, Haijun Zhang, Jianghong Ma, Jicong Fan, and Zhao Zhang. 2023. Fcboost-net: A generative network for synthesizing multiple collocated outfits via fashion compatibility boosting. InProceedings of the 31st ACM International Conference on Multimedia. 7881–7889

    2023