Pith. sign in

REVIEW 4 major objections 6 minor 2 cited by

Imagine and Seek: Improving Composed Image Retrieval with an Imagined Proxy

T0 review · 4 major / 6 minor · reviewed 2026-08-12 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read This paper claims that zero-shot composed image retrieval can be improved without any training by generating an "imagined proxy" image—a synthetic picture of the query after the caption's edit—and merging its visual features with the…

desk verdict A training-free 'imagined proxy' improves ZS-CIR across three datasets; the core idea is new, the ablations are honest, but the paper needs a cheap-control experiment and implementation details before fully trusting the causal claim. read the letter →

arxiv 2411.16752 v2 pith:VSOPB4B2 submitted 2024-11-24 cs.CV

You Li , Fan Ma , Yi Yang This is my paper · ORCID
classification cs.CV
keywords composedimageretrievalzero-shotimaginedproxytraining-freemethodlargelanguagemodelcontrollablegenerationsemanticperturbationbalancing
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper tries to establish that zero-shot composed image retrieval—searching for images that match a query image plus a relative caption—can be improved without any training by first imagining what the target looks like. Its proposal, IP-CIR, generates a synthetic "proxy image" that applies the caption's edits to the query image, then uses that proxy's visual features to supplement the text features that existing methods rely on. The authors argue that text-only retrieval misses fine-grained visual detail such as color, style, and spatial layout, and they show that composing the proxy feature with the query feature and a text semantic perturbation, plus a balancing metric between text and proxy similarities, lifts several strong baselines on CIRR, CIRCO, and FashionIQ. A sympathetic reader would care because this offers a training-free, plug-and-play route to bring image-side information into a retrieval pipeline that is otherwise text-centric.

What carries the argument

The load-bearing mechanism is the imagined proxy and its feature composition. The proxy is a generated image meant to be the query image after the caption's edits, produced by an LLM-reasoned layout (object descriptions, bounding boxes, and per-instance reference modality) rendered by a controllable multi-instance generator that accepts reference-image guidance. The central identity is the robust proxy feature $f_{RP}=f_p+\frac{\max(f_p)}{\max(f_q)}f_q+\frac{\max(f_p)}{\max(f_s)}f_s$, which rescales the query and semantic-perturbation features to the proxy feature's magnitude, plus the balancing similarity $S_f=\lambda S_t+(1-\lambda)S_tS_p$, whose product term requires a candidate to score well in both text and proxy evidence. This composition does the work: the ablation in Table 3 shows that raw proxy features alone can slightly hurt CIRR Recall@1 (36.15 to 36.02), while the composed robust feature and balance metric convert the proxy into consistent gains.

What would settle it

Take a set of query/caption pairs in which the caption alters exactly one attribute (say, color) and check whether the generated proxy preserves all other attributes; if many proxies misapply the edit, and the full method's Recall@1 never beats the text-only baseline for any setting of the balance parameter, the central claim is refuted.

Watch

Extended reading notes

Core claim

The central claim is that an imagined proxy image carries visual information—instance attributes, style, spatial relationships—that CLIP text features omit, and that this information can be recovered and used for retrieval without any training. IP-CIR first uses an LLM, given a BLIP2 caption of the query image and the relative caption, to generate a layout of objects with bounding boxes and per-instance reference modalities (whether appearance should come from the query image, the text, or both). A controllable generator then renders this layout, using the query image as reference, to produce the proxy. The proxy's CLIP feature $f_p$ is merged with the query feature $f_q$ and a semantic perturbation $f_s=f_t-f_o$ (the difference between LLM-inferred target-caption and original-caption features) into a robust feature $f_{RP}=f_p+\frac{\max(f_p)}{\max(f_q)}f_q+\frac{\max(f_p)}{\max(f_s)}f_s$, and retrieval uses the balanced score $S_f=\lambda S_t+(1-\lambda)S_tS_p$, where $S_t$ is the text-side similarity and $S_p$ the proxy-side similarity. As reported in Table 1, this raises CIRR Recall@5 from 66.39 to 70.07 and Recall@10 from 77.25 to 80.00 for LDRE with a ViT-G/14 backbone, and improves CIRCO mAP@10 from 32.24 to 34.26 and FashionIQ average Recall@10 from 45.11 to 45.74.

Load-bearing premise

The whole method rests on the assumption that the imagined proxy image faithfully keeps the query image's key attributes while applying the caption's changes; if the generated proxy drifts from the query, the merged features inherit that drift and retrieval suffers.

Editorial extensions

If this is right

  • Any existing text-based zero-shot CIR method can be wrapped by IP-CIR without retraining; the paper demonstrates improvements when it is added to both LDRE and LinCIR across CLIP ViT-L and ViT-G backbones.
  • The number of imagined proxies is a practical dial: retrieval accuracy rises with more proxies and the gain saturates, so users can trade compute for accuracy.
  • The per-dataset balance weight λ must be tuned (0.0 for CIRR, 0.3 for CIRCO, 0.8 for FashionIQ), and the paper notes this reflects how well the generated proxies match each dataset's captions.
  • Proxy quality is the pivot: when generation is imperfect (e.g., FashionIQ logos and white backgrounds), proxy evidence is less reliable and the method relies on a larger λ, so improving controllable generation should translate directly into retrieval gains.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • One could extend the same proxy-plus-balance recipe to other relative-to-image retrieval tasks, such as fashion recommendation or video retrieval, since nothing in the pipeline depends on the three benchmarks used.
  • The sensitivity to λ suggests an automatic, per-query weighting scheme—for instance, estimating proxy reliability from the agreement among the five generated proxies—could remove the main manual tuning step.
  • Failures on attributes that are hard to generate (logos, pure white backgrounds) point to generation quality rather than fusion as the bottleneck, implying that better controllable generators would improve retrieval even without changing the retrieval method.
  • The multiplicative balance $S_tS_p$ acts as a soft AND over the two modalities; comparing it with rank-based or learned fusion on the same features would test whether the specific product form is essential or just a convenient choice.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 6 minor

Summary. This paper proposes IP-CIR, a training-free method for zero-shot composed image retrieval (ZS-CIR). The method first uses BLIP-2 captions and an LLM to infer an object layout from the query image and relative caption, then uses MIGC++ conditioned on the query image to generate multiple 'proxy' images. A robust retrieval feature fRP is formed by adding the query image feature and a semantic perturbation (difference between LLM-generated target captions and query captions) to the proxy feature, and a balance metric combines text-based similarity and proxy-based similarity via Sf = λSt + (1−λ)StSp. The method is evaluated on CIRR, CIRCO, and FashionIQ when added to LDRE and LinCIR baselines, with reported improvements and state-of-the-art results on CIRR. The paper claims the imagined proxy provides fine-grained visual information that text-only features miss.

Significance. The idea of using conditional image generation to create a proxy retrieval feature is original, and the plug-and-play nature of the method is attractive. The paper uses standard datasets and baselines, and the qualitative examples show plausible cases where the proxy captures attributes that text-only retrieval misses. However, the current experiments do not isolate the proxy's causal contribution: the ablations in Table 3 show that direct proxy similarity alone does not help, and the full gain appears only with the robust feature and balance metric, both of which include query-image and semantic-perturbation terms. Without a control that removes the proxy feature, the claimed central mechanism remains unverified. The reported CIRR headline metric is also misstated in the abstract. These issues are fixable with additional experiments and corrections, so the paper's potential contribution is meaningful but not yet established.

major comments (4)
  1. [§4.4, Table 3, Eq. (1)] The ablation does not isolate the contribution of the imagined proxy. The row 'PI' (direct proxy similarity) shows a small decrease for CIRR R@1 (-0.13) and CIRCO mAP@5 (-0.06), whereas the full gain appears only after adding the robust proxy feature fRP (which always includes fp) and the balance metric Sb = St*Sp. Because every ablation with RP includes fp, there is no condition that removes fp while keeping fq and fs; the improvement attributed to the proxy could instead come from the fq + fs terms and from the product re-ranking, which is exact when λ = 0 for CIRR. To support the novelty claim, the authors should add controls that replace fp with (a) the query image's own CLIP feature, (b) an image generated from the text layout without query-image conditioning, or (c) a random image feature; if these controls obtain similar gains, the expensive MIGC++ generation step is not the causal ingredient.
  2. [Abstract and Introduction] The reported CIRR result is misstated. The abstract and the Introduction both claim a Recall@K of 70.07 at K=10 on CIRR, but Table 1 shows that for LDRE+IP-CIR with ViT-G/14 the Recall@10 is 80.00 and 70.07 is the Recall@5 value. This affects the headline claim and must be corrected.
  3. [§3.3 and §4.1] Several implementation details required for reproducibility are missing. The paper states that five proxy images are generated per query but does not specify how their features are ensembled into a single fp (e.g., element-wise averaging, max pooling, or selection). It also does not provide the LLM prompt used to obtain the target captions ft that define the semantic perturbation fs = ft − fo, nor the number of such captions. Without these details, the construction of fRP cannot be reproduced.
  4. [General experimental reporting] No error bars or significance tests are reported. Several improvements are small relative to the variation expected in retrieval metrics, e.g., LinCIR+IP-CIR gains of +0.22 R@1 on CIRR and −0.30 R@10 on FashionIQ Toptee. The authors should report standard deviations over multiple runs or at least over multiple seeds, or justify that the observed differences are stable.
minor comments (6)
  1. [Throughout] There are typos and inconsistencies, such as 'Implement Details' for 'Implementation Details' and inconsistent use of 'balance metric' versus 'balanced metric'.
  2. [Eq. (1)] The notation max(fp) is ambiguous: it is unclear whether the maximum is taken over the feature dimension or over the entire vector, and the rationale for scaling by the ratio of maxima is not explained.
  3. [Appendix and main text figures] The appendix restarts figure numbering with 'Figure 1', which conflicts with the main text's Figure 1; renumber the appendix figures.
  4. [§3.2] The prompt template for the LLM layout generation is only partially given ('Given an image of {caption}, we show {rule}'); the full instruction, including the format for bbox coordinates and reference modalities, should be included.
  5. [§4.1] The description of the CIRCO test set evaluation is ambiguous; state explicitly whether the official 800-query test split is used and whether the standard exclusion of the query image is applied.
  6. [Figure 3] The text says five proxies are generated, but the figure shows only two; clarify that these are example proxy images rather than the full ensemble.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the method is a heuristic pipeline whose retrieval claim is benchmarked externally; self-cited MIGC/MIGC++ generator is not used to justify the retrieval result.

full rationale

IP-CIR contains no derivation step that reduces by construction to its own inputs. Equation (1) fuses proxy features fp with query features fq and semantic perturbation fs using max-ratio normalization, and Eq. (2) combines baseline and proxy similarities as Sf = λSt + (1−λ)StSp. The per-dataset λ values (0.3 for CIRCO, 0.0 for CIRR, 0.8 for FashionIQ) are ordinary validation-set hyperparameter choices, not fitted parameters renamed as predictions. The only load-bearing self-citations are to MIGC [47] and MIGC++ [48], which supply the proxy-image generator; these are externally published, benchmarked tools, and the paper does not invoke their results to prove the retrieval improvement. Table 3 does show that raw proxy features alone can slightly hurt on some metrics (CIRR R@1 36.15 to 36.02; CIRCO mAP@5 31.12 to 31.06), with gains appearing after adding query and semantic-perturbation terms; this is an ablation-design concern about isolating the proxy contribution, not a circular argument. The appendix limitation 'Sensitive to hyperparameters' further confirms that λ is a tuned trade-off rather than a hidden circular input. Because the method is evaluated against public ZS-CIR baselines on CIRR, CIRCO, and FashionIQ, the central claim is externally falsifiable and independent of the paper's own definitions.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

The method is an empirical pipeline that depends on three tunable quantities (λ, proxy count, feature weights) and on several domain assumptions about proxy image quality and the meaningfulness of semantic perturbation in CLIP space. No new theoretical entities are introduced.

free parameters (3)
  • λ (balance weight) = CIRCO: 0.3, CIRR: 0.0, FashionIQ: 0.8
    Dataset-specific weight in Eq. 2; tuned on validation (Fig.7). For CIRR λ=0, which discards text-only similarity and uses the product of text and proxy similarities.
  • number of proxy images = 5
    Chosen as a trade-off; performance increases with number but saturates (Fig.8).
  • feature weights in Eq. 1 = 1,1,1 by default, adjustable
    Paper states 'weights of each component can be adjusted' (Section 3.3), but no values or tuning details are given.
assumptions (4)
  • domain assumption CLIP image features of proxy images are informative for retrieval similarity with target images
    The method relies on fp being comparable to target image CLIP features; ablation shows this is not always true (PI alone can reduce performance).
  • domain assumption MIGC++ with ELITE can generate proxy images that reflect the relative caption's edits while preserving query image identity
    The whole proxy-generation step assumes controllable generation succeeds; quality varies by dataset (Appendix A.2 notes FashionIQ proxies struggle with logos and white backgrounds).
  • domain assumption Semantic perturbation fs = ft - fo in CLIP space is a meaningful edit direction
    Borrowed from LDRE [43]; the paper applies it to image features without validation.
  • ad hoc to paper The balancing metric Sf = λSt + (1-λ)StSp is a better combination than simple averaging
    Motivated by a toy example about extreme similarities; no theoretical derivation or empirical comparison to other fusion methods.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Imagine and Seek: Improving Composed Image Retrieval with an Imagined Proxy." pith.science (2026). https://pith.science/paper/VSOPB4B2

@misc{pith2026241116752,
  author       = {Pith},
  title        = {Pith review of: Imagine and Seek: Improving Composed Image Retrieval with an Imagined Proxy},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/VSOPB4B2}},
  note         = {Machine review of arXiv:2411.16752}
}
read the original abstract

The Zero-shot Composed Image Retrieval (ZSCIR) requires retrieving images that match the query image and the relative captions. Current methods focus on projecting the query image into the text feature space, subsequently combining them with features of query texts for retrieval. However, retrieving images only with the text features cannot guarantee detailed alignment due to the natural gap between images and text. In this paper, we introduce Imagined Proxy for CIR (IP-CIR), a training-free method that creates a proxy image aligned with the query image and text description, enhancing query representation in the retrieval process. We first leverage the large language model's generalization capability to generate an image layout, and then apply both the query text and image for conditional generation. The robust query features are enhanced by merging the proxy image, query image, and text semantic perturbation. Our newly proposed balancing metric integrates text-based and proxy retrieval similarities, allowing for more accurate retrieval of the target image while incorporating image-side information into the process. Experiments on three public datasets demonstrate that our method significantly improves retrieval performances. We achieve state-of-the-art (SOTA) results on the CIRR dataset with a Recall@K of 70.07 at K=10. Additionally, we achieved an improvement in Recall@10 on the FashionIQ dataset, rising from 45.11 to 45.74, and improved the baseline performance in CIRCO with a mAPK@10 score, increasing from 32.24 to 34.26.

Figures

Figures reproduced from arXiv: 2411.16752 by the authors.

Figure 1
Figure 1. Comparison of composed image retrieval between text-only retrieval and our methods. Traditional methods per￾form retrieval only in the text space, where CLIP text features may overlook some important semantic information. In contrast, our approach generates imagined proxy features, providing additional information that is often overlooked in text-only retrieval, thereby improving retrieval accuracy. 49] and image re… view at source ↗
Figure 2
Figure 2. Overview of our method. (a) represents our imagined proxy generation process. We use LLM to analyze the BLIP2-generated query image captions and the relative captions and infer the proxy layout. We then use the controllable generator to imagine the proxy images. (b) represents our process of constructing a robust proxy feature, and balancing the text and proxy similarities. We integrate proxy features, query image f… view at source ↗
Figure 3
Figure 3. Qualitative results of our method. We conducted experiments on the CIRCO validation dataset to observe in which cases our method improves retrieval results. In the blue section on the left, we display the query information used for retrieval, the ground truth target image, and our generated proxy image features. ‘Query’ represents the input query image, ‘Caption’ represents the relative text, and in the ’Proxy’ sect… view at source ↗
Figures from the paper (10 more)
Figure 4
Figure 4. Figure 4: Qualitative results of our method. We present the result on CIRCO validation dataset with top-1 retrieval results. Baseline methods in text-based retrieval overlook certain attributes and details in complex related texts, resulting in suboptimal re￾trieval outcomes. In…
Figure 5
Figure 5. Figure 5: Ablation results on the composition of robust proxy features in the CIRCO dataset. pi indicates proxy features, si rep￾resents semantic perturbation, and qi indicates the query features. 4.3. Qualitative Results We conducted experiments on the CIRCO validation dataset …
Figure 6
Figure 6. Figure 6: Visualize of Ablation Result. We present the ablation result on CIRCO dataset. trieves a more suitable image in the top-1 result. Analysis of Experimental Results on TOP-5 Retrieval. From the visualized results in Fig.3, we can draw the fol￾lowing conclusions: (1) Our …
Figure 7
Figure 7. Figure 7: The impact of weight λ in CIRCO, CIRR, and FashionIQ. As λ decreases, the proportion of proxy information increases. Proxy) indicates integrating the proxy image into more ro￾bust proxy features. BM (Balance Metric) refers to using the balance metric. The results from …
Figure 8
Figure 8. Figure 8: Parameter Analysis. The effect of varying numbers of proxy images on performance for the CIRCO dataset. Effects of the weight λ. We analyze the λ used in Eq. (2). The experiments revealed that differences in dataset prop￾erties, result in different optimal values for λ…
Figure 1
Figure 1. Figure 1: Ablation results on the composition of robust proxy features in the CIRCO dataset. pi indicates proxy features, si rep￾resents semantic perturbation, and qi indicates the query features. A. More Analysis In this section, we will conduct a deeper analysis of the ex￾peri…
Figure 2
Figure 2. Figure 2: Ablation result on the Robust Proxy. We present the visualization result of using different compositions (Qi represents only using query image, Si represents only semantic perturbation, and Pi represents only using the proxy image) of features in the robust proxy. has …
Figure 3
Figure 3. Figure 3: Qualitative results on CIRCO dataset. We show more improvement in top-1 retrieval results in the CIRCO dataset. 2 [PITH_FULL_IMAGE:figures/full_fig_p012_3.png]
Figure 4
Figure 4. Figure 4: Qualitative results on CIRR dataset. We show more improvement in top-1 retrieval results in the CIRR dataset. ple, in the first column, the car we retrieved is yellow. In the second column, the background of our result has sand instead of water. In the third column, wi…
Figure 5
Figure 5. Figure 5: Qualitative results on FashionIQ dataset. We show more improvement in top-1 retrieval results in the FashionIQ dataset. 4 [PITH_FULL_IMAGE:figures/full_fig_p014_5.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 2 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. AnySynth: Harnessing the Power of Image Synthetic Data Generation for Generalized Vision-Language Tasks

    cs.CV 2024-11 conditional novelty 5.0 of 10

    AnySynth is a single synthetic-data pipeline that produces layouts, images, and annotations for multiple vision tasks, and its data improves benchmark scores by small but consistent margins.

  2. MCoT-RE: Multi-Faceted Chain-of-Thought and Re-Ranking for Training-Free Zero-Shot Composed Image Retrieval

    cs.CV 2025-07 conditional novelty 4.0 of 10

    A training-free composed image retrieval method uses multi-faceted chain-of-thought prompting to produce modification-focused and integration-focused captions, then filters and re-ranks CLIP candidates with a weighted fusion.

Reference graph

Works this paper leans on

49 extracted references · 31 canonical work pages · cited by 2 Pith papers

  1. [1]

    isearle: Improving textual inversion for zero-shot composed image retrieval

    Lorenzo Agnolucci, Alberto Baldrati, Marco Bertini, and Alberto Del Bimbo. isearle: Improving textual inversion for zero-shot composed image retrieval. arXiv preprint arXiv:2405.02951, 2024. 1

    arXiv 2024

  2. [2]

    Qwen technical report

    Jinze Bai, Shuai Bai, Yunfei Chu, Zeyu Cui, Kai Dang, Xi- aodong Deng, Yang Fan, Wenbin Ge, Yu Han, Fei Huang, Binyuan Hui, Luo Ji, Mei Li, Junyang Lin, Runji Lin, Day- iheng Liu, Gao Liu, Chengqiang Lu, Keming Lu, Jianxin Ma, Rui Men, Xingzhang Ren, Xuancheng Ren, Chuanqi Tan, Sinan Tan, Jianhong Tu, Peng Wang, Shijie Wang, Wei Wang, Shengguang Wu, Benfe...

    arXiv 2023

  3. [3]

    ediff-i: Text-to-image diffusion models with ensemble of expert denoisers

    Yogesh Balaji, Seungjun Nah, Xun Huang, Arash Vahdat, Jiaming Song, Qinsheng Zhang, Karsten Kreis, Miika Ait- tala, Timo Aila, Samuli Laine, Bryan Catanzaro, Tero Kar- ras, and Ming-Yu Liu. ediff-i: Text-to-image diffusion models with ensemble of expert denoisers. arXiv preprint arXiv:2211.01324, 2022. 1

    arXiv 2022

  4. [4]

    Effective conditioned and composed im- age retrieval combining clip-based features

    Alberto Baldrati, Marco Bertini, Tiberio Uricchio, and Al- berto Del Bimbo. Effective conditioned and composed im- age retrieval combining clip-based features. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pat- tern Recognition, pages 21466–21474, 2022. 1

    work page 2022

  5. [5]

    Conditioned and composed image retrieval combining and partially fine-tuning clip-based features

    Alberto Baldrati, Marco Bertini, Tiberio Uricchio, and Al- berto Del Bimbo. Conditioned and composed image retrieval combining and partially fine-tuning clip-based features. In Proceedings of the IEEE/CVF Conference on Computer Vi- sion and Pattern Recognition, pages 4959–4968, 2022. 1

    work page 2022

  6. [6]

    Zero-shot composed image retrieval with textual inversion, 2023

    Alberto Baldrati, Lorenzo Agnolucci, Marco Bertini, and Al- berto Del Bimbo. Zero-shot composed image retrieval with textual inversion, 2023. 1, 2, 5

    work page 2023

  7. [7]

    Zero-shot composed image retrieval with textual inversion

    Alberto Baldrati, Lorenzo Agnolucci, Marco Bertini, and Al- berto Del Bimbo. Zero-shot composed image retrieval with textual inversion. In Proceedings of the IEEE/CVF Interna- tional Conference on Computer Vision, pages 15338–15347,

  8. [8]

    Attend-and-excite: Attention-based se- mantic guidance for text-to-image diffusion models, 2023

    Hila Chefer, Yuval Alaluf, Yael Vinker, Lior Wolf, and Daniel Cohen-Or. Attend-and-excite: Attention-based se- mantic guidance for text-to-image diffusion models, 2023. 2

    2023

Show all 49 references
  1. [9]

    Anydoor: Zero-shot object-level im- age customization

    Xi Chen, Lianghua Huang, Yu Liu, Yujun Shen, Deli Zhao, and Hengshuang Zhao. Anydoor: Zero-shot object-level im- age customization. arXiv preprint arXiv:2307.09481, 2023. 2

    2023 arXiv

  2. [10]

    Learning joint visual se- mantic matching embeddings for language-guided retrieval

    Yanbei Chen and Loris Bazzani. Learning joint visual se- mantic matching embeddings for language-guided retrieval. page 136–152, Berlin, Heidelberg, 2020. Springer-Verlag. 1

    2020

  3. [11]

    Image search with text feedback by visiolinguistic attention learn- ing

    Yanbei Chen, Shaogang Gong, and Loris Bazzani. Image search with text feedback by visiolinguistic attention learn- ing. In 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 2998–3008, 2020

    2020

  4. [12]

    Composed image retrieval with text feedback via multi-grained uncertainty regularization, 2024

    Yiyang Chen, Zhedong Zheng, Wei Ji, Leigang Qu, and Tat- Seng Chua. Composed image retrieval with text feedback via multi-grained uncertainty regularization, 2024

    2024

  5. [13]

    Improved probabilistic image-text repre- sentations, 2024

    Sanghyuk Chun. Improved probabilistic image-text repre- sentations, 2024

    2024

  6. [14]

    Artemis: Attention-based re- trieval with text-explicit matching and implicit similarity, 2022

    Ginger Delmas, Rafael Sampaio de Rezende, Gabriela Csurka, and Diane Larlus. Artemis: Attention-based re- trieval with text-explicit matching and implicit similarity, 2022

    2022

  7. [15]

    Modality-agnostic attention fusion for vi- sual search with text feedback, 2020

    Eric Dodds, Jack Culpepper, Simao Herdade, Yang Zhang, and Kofi Boakye. Modality-agnostic attention fusion for vi- sual search with text feedback, 2020. 1

    2020

  8. [16]

    Hierarchical text-conditional image generation with clip latents, 2022

    Aditya Ramesh et al. Hierarchical text-conditional image generation with clip latents, 2022. 1

    2022

  9. [17]

    Language-only training of zero- shot composed image retrieval

    Geonmo Gu, Sanghyuk Chun, Wonjae Kim, , Yoohoon Kang, and Sangdoo Yun. Language-only training of zero- shot composed image retrieval. In Conference on Computer Vision and Pattern Recognition (CVPR), 2024. 1, 2, 6

    2024

  10. [18]

    Discriminative learning of open-vocabulary object retrieval and localization by neg- ative phrase augmentation

    Ryota Hinami and Shin’ichi Satoh. Discriminative learning of open-vocabulary object retrieval and localization by neg- ative phrase augmentation. In Proceedings of the 2018 Con- ference on Empirical Methods in Natural Language Process- ing, pages 2605–2615, Brussels, Belgium, ...

    2018

  11. [19]

    Denoising dif- fusion probabilistic models

    Jonathan Ho, Ajay Jain, and Pieter Abbeel. Denoising dif- fusion probabilistic models. Advances in neural information processing systems, 33:6840–6851, 2020. 1, 2

    2020

  12. [20]

    Vision-by-language for training-free compositional image retrieval

    Shyamgopal Karthik, Karsten Roth, Massimiliano Mancini, and Zeynep Akata. Vision-by-language for training-free compositional image retrieval. International Conference on Learning Representations (ICLR), 2024. 2, 6

    2024

  13. [21]

    Jooyeon Kim, Eulrang Cho, Sehyung Kim, and Hyunwoo J. Kim. Retrieval-augmented open-vocabulary object detec- tion, 2024. 1

    2024

  14. [22]

    BLIP-2: bootstrapping language-image pre-training with frozen image encoders and large language models

    Junnan Li, Dongxu Li, Silvio Savarese, and Steven Hoi. BLIP-2: bootstrapping language-image pre-training with frozen image encoders and large language models. In ICML,

  15. [23]

    Gligen: Open-set grounded text-to-image generation

    Yuheng Li, Haotian Liu, Qingyang Wu, Fangzhou Mu, Jian- wei Yang, Jianfeng Gao, Chunyuan Li, and Yong Jae Lee. Gligen: Open-set grounded text-to-image generation. CVPR,

  16. [24]

    Anysynth: Harnessing the power of image synthetic data generation for generalized vision-language tasks

    You Li, Fan Ma, and Yi Yang. Anysynth: Harnessing the power of image synthetic data generation for generalized vision-language tasks. arXiv preprint arXiv:2411.16749 ,

    arXiv

  17. [25]

    Caphuman: Capture your moments in parallel uni- verses

    Chao Liang, Fan Ma, Linchao Zhu, Yingying Deng, and Yi Yang. Caphuman: Capture your moments in parallel uni- verses. In CVPR, pages 6400–6409, 2024. 1

    2024

  18. [26]

    Lawrence Zitnick, and Piotr Doll ´ar

    Tsung-Yi Lin, Michael Maire, Serge Belongie, Lubomir Bourdev, Ross Girshick, James Hays, Pietro Perona, Deva Ramanan, C. Lawrence Zitnick, and Piotr Doll ´ar. Microsoft coco: Common objects in context, 2015. 6 9

    2015

  19. [27]

    Image retrieval on real-life images with pre-trained vision-and-language models

    Zheyuan Liu, Cristian Rodriguez-Opazo, Damien Teney, and Stephen Gould. 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, 2021. 1, 2, 5

    2021

  20. [28]

    Thinking fast and slow: Effi- cient text-to-visual retrieval with transformers, 2021

    Antoine Miech, Jean-Baptiste Alayrac, Ivan Laptev, Josef Sivic, and Andrew Zisserman. Thinking fast and slow: Effi- cient text-to-visual retrieval with transformers, 2021. 1

    2021

  21. [29]

    Glide: Towards photorealistic image generation and editing with text-guided diffusion models, 2022

    Alex Nichol, Prafulla Dhariwal, Aditya Ramesh, Pranav Shyam, Pamela Mishkin, Bob McGrew, Ilya Sutskever, and Mark Chen. Glide: Towards photorealistic image generation and editing with text-guided diffusion models, 2022. 1

    2022

  22. [30]

    Gpt-4 technical report, 2023

    OpenAI. Gpt-4 technical report, 2023. 2, 3, 4

    2023

  23. [31]

    Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, Gretchen Krueger, and Ilya Sutskever

    Alec Radford, Jong Wook Kim, Chris Hallacy, A. Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, Gretchen Krueger, and Ilya Sutskever. Learning transferable visual models from nat- ural language supervision. In ICML, 2021. 2

    2021

  24. [32]

    High-resolution image syn- thesis with latent diffusion models, 2021

    Robin Rombach, Andreas Blattmann, Dominik Lorenz, Patrick Esser, and Bj¨orn Ommer. High-resolution image syn- thesis with latent diffusion models, 2021. 1, 2

    2021

  25. [33]

    Denton, Seyed Kamyar Seyed Ghasemipour, Burcu Karagol Ayan, Seyedeh Sara Mahdavi, Raphael Gontijo Lopes, Tim Salimans, Jonathan Ho, David Fleet, and Mohammad Norouzi

    Chitwan Saharia, William Chan, Saurabh Saxena, Lala Li, Jay Whang, Emily L. Denton, Seyed Kamyar Seyed Ghasemipour, Burcu Karagol Ayan, Seyedeh Sara Mahdavi, Raphael Gontijo Lopes, Tim Salimans, Jonathan Ho, David Fleet, and Mohammad Norouzi. Photorealistic text-to-image diffu...

    2022

  26. [34]

    Pic2word: Mapping pictures to words for zero-shot composed image retrieval

    Kuniaki Saito, Kihyuk Sohn, Xiang Zhang, Chun-Liang Li, Chen-Yu Lee, Kate Saenko, and Tomas Pfister. Pic2word: Mapping pictures to words for zero-shot composed image retrieval. CVPR, 2023. 1, 2

    2023

  27. [35]

    Denois- ing diffusion implicit models

    Jiaming Song, Chenlin Meng, and Stefano Ermon. Denois- ing diffusion implicit models. arXiv:2010.02502, 2020. 1, 2

    2010 arXiv

  28. [36]

    A corpus for reasoning about natural language grounded in photographs, 2019

    Alane Suhr, Stephanie Zhou, Ally Zhang, Iris Zhang, Huajun Bai, and Yoav Artzi. A corpus for reasoning about natural language grounded in photographs, 2019. 5

    2019

  29. [37]

    Instancediffusion: Instance-level control for image generation, 2024

    Xudong Wang, Trevor Darrell, Sai Saketh Rambhatla, Rohit Girdhar, and Ishan Misra. Instancediffusion: Instance-level control for image generation, 2024. 2

    2024

  30. [38]

    Re-identification = retrieval + verification: Back to essence and forward with a new metric

    Zheng Wang, Xin Yuan, Toshihiko Yamasaki, Yutian Lin, Xin Xu, and Wenjun Zeng. Re-identification = retrieval + verification: Back to essence and forward with a new metric. arXiv preprint arXiv:2011.11506, 2020. 1

    2011 arXiv

  31. [39]

    Elite: Encoding visual con- cepts into textual embeddings for customized text-to-image generation

    Yuxiang Wei, Yabo Zhang, Zhilong Ji, Jinfeng Bai, Lei Zhang, and Wangmeng Zuo. Elite: Encoding visual con- cepts into textual embeddings for customized text-to-image generation. arXiv preprint arXiv:2302.13848, 2023. 2, 4

    2023 arXiv

  32. [40]

    The fashion iq dataset: Retrieving images by combining side information and relative natural language feedback

    Hui Wu, Yupeng Gao, Xiaoxiao Guo, Ziad Al-Halah, Steven Rennie, Kristen Grauman, and Rogerio Feris. The fashion iq dataset: Retrieving images by combining side information and relative natural language feedback. CVPR, 2021. 1, 2, 5, 6

    2021

  33. [41]

    Boxdiff: Text-to-image synthesis with training-free box-constrained diffusion

    Jinheng Xie, Yuexiang Li, Yawen Huang, Haozhe Liu, Wen- tian Zhang, Yefeng Zheng, and Mike Zheng Shou. Boxdiff: Text-to-image synthesis with training-free box-constrained diffusion. arXiv preprint arXiv:2307.10816, 2023. 2

    2023 arXiv

  34. [42]

    Reco: Region-controlled text-to- image generation

    Zhengyuan Yang, Jianfeng Wang, Zhe Gan, Linjie Li, Kevin Lin, Chenfei Wu, Nan Duan, Zicheng Liu, Ce Liu, Michael Zeng, and Lijuan Wang. Reco: Region-controlled text-to- image generation. In CVPR, 2023. 2

    2023

  35. [43]

    Ldre: Llm-based divergent reasoning and ensemble for zero-shot composed image re- trieval

    Zhenyu Yang, Dizhan Xue, Shengsheng Qian, Weiming Dong, and Changsheng Xu. Ldre: Llm-based divergent reasoning and ensemble for zero-shot composed image re- trieval. In Proceedings of the 47th International ACM SI- GIR Conference on Research and Development in Informa- tion Re...

    2024

  36. [44]

    Magi- cLens: Self-supervised image retrieval with open-ended in- structions

    Kai Zhang, Yi Luan, Hexiang Hu, Kenton Lee, Siyuan Qiao, Wenhu Chen, Yu Su, and Ming-Wei Chang. Magi- cLens: Self-supervised image retrieval with open-ended in- structions. In Proceedings of the 41st International Con- ference on Machine Learning, pages 59403–59420. PMLR,

  37. [45]

    Hauptmann

    Liang Zheng, Yi Yang, and Alexander G. Hauptmann. Per- son re-identification: Past, present and future, 2016. 1

    2016

  38. [46]

    Pyramid diffusion models for low-light image enhancement

    Dewei Zhou, Zongxin Yang, and Yi Yang. Pyramid diffusion models for low-light image enhancement. arXiv preprint arXiv:2305.10028, 2023. 1

    2023 arXiv

  39. [47]

    Migc: Multi-instance generation controller for text-to-image synthesis

    Dewei Zhou, You Li, Fan Ma, Zongxin Yang, and Yi Yang. Migc: Multi-instance generation controller for text-to-image synthesis. 2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , pages 6818–6828, 2024. 2, 6

    2024

  40. [48]

    Migc++: Advanced multi-instance generation controller for image synthesis

    Dewei Zhou, You Li, Fan Ma, Zongxin Yang, and Yi Yang. Migc++: Advanced multi-instance generation controller for image synthesis. ArXiv, abs/2407.02329, 2024. 2, 4

    2024 arXiv

  41. [49]

    3dis: Depth-driven decoupled instance synthesis for text-to-image generation

    Dewei Zhou, Ji Xie, Zongxin Yang, and Yi Yang. 3dis: Depth-driven decoupled instance synthesis for text-to-image generation. ArXiv, abs/2410.12669, 2024. 1 10 Appendix Figure 1. Ablation results on the composition of robust proxy features in the CIRCO dataset. pi indicates pro...

    2024

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.