Pith. sign in

REVIEW 3 major objections 4 minor 45 references

Harlequin: Color-driven Generation of Synthetic Data for Referring Expression Comprehension

T0 review · 3 major / 4 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 Fully synthetic color-swapped training data matches or improves referring-expression models on real benchmarks.

desk verdict A believable synthetic-data result for REC, with an unmeasured generation-fidelity link and no real-data pre-training baseline; the core finding is solid but the mechanism is not fully pinned down. read the letter →

arxiv 2411.14807 v1 pith:AM5LX225 submitted 2024-11-22 cs.CV cs.CLcs.LG

Luca Parolari , Elena Izzo , Lamberto Ballan This is my paper · ORCID
classification cs.CVcs.CLcs.LG
keywords syntheticdatagenerationreferringexpressioncomprehensionvisualgroundingtext-to-imagediffusioncolorattributeaugmentationGLIGENpre-trainingFlickr30kEntities
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 Referring Expression Comprehension (REC), locating the object named by a natural-language phrase, can be trained with synthetic data that requires no human annotation. The authors build Harlequin, a dataset of 286,948 images and 1,093,181 queries, by taking color-bearing referring expressions from Flickr30k Entities, swapping the color token for one from a 12-color vocabulary, and rendering a new image with GLIGEN conditioned on the edited caption and box-level entity descriptions. Pre-training TransVG, VLTVG, and LGR-NET on Harlequin before fine-tuning on RefCOCO, RefCOCO+, and RefCOCOg gives accuracy equal to or better than training from scratch on real data, with the largest gains on RefCOCOg for LGR-NET (+4.11 on the test split). If this holds, expensive manual annotation is no longer a hard bottleneck for REC, and synthetic benchmarks of arbitrary size and controlled complexity become feasible.

What carries the argument

The load-bearing mechanism has two parts. The Annotation Generation Engine is a function $\varphi: A \to A$ that edits an annotation by replacing the color attribute in the chosen referring expression with a color sampled from a fixed vocabulary, leaving the bounding boxes and the other referring expressions untouched. The Image Generation Engine is a function $\psi: A \to I$ implemented with GLIGEN, a grounded text-to-image model that renders an image from the edited caption together with the box-plus-text entities. What makes the pair work is that the textual change is chosen to preserve layout while forcing a visible appearance change, so the synthetic labels attach a discriminative color signal to a known location.

What would settle it

Take a random sample of Harlequin images and run an attribute-aware detector to verify, for each query, that the object inside the bounding box carries the swapped color and that the unswitched context objects are still present. If a substantial share of images fails this fidelity check while the fine-tuning gains remain, then accurate rendering is not the active ingredient; if the gains disappear when Harlequin images are replaced by unrelated images paired with the same queries, rendering fidelity is the mechanism.

Watch

Extended reading notes

Core claim

On the paper's own terms, the discovery is that a single attribute, color, is enough to generate synthetic REC supervision that transfers to real benchmarks. The pipeline treats each seed annotation as a caption plus a set of (referring expression, bounding box) pairs; it replaces the color in one referring expression and the matching caption span, keeps every box and every other expression fixed, and asks GLIGEN to render an image consistent with both the edited caption and the entity boxes. The resulting dataset, Harlequin, contains over one million queries, and pre-training on it improves or matches the from-scratch baselines for all three models and all three benchmark families; on color-only test subsets, TransVG improves by up to +5.15 on RefCOCO val. The authors conclude that the artificial character of the data is overcome when the generative model is controlled tightly, and that fully automatic annotation for REC is achievable.

Load-bearing premise

The result stands on GLIGEN faithfully painting the swapped color onto the referred object inside its box: if the generator often ignores the color or location guidance, Harlequin's labels are noisy and the measured gains could come from corpus size rather than accurate synthetic supervision.

Editorial extensions

If this is right

  • Pre-training on Harlequin works as a drop-in initialization: TransVG, VLTVG, and LGR-NET need no architectural changes to benefit from it.
  • Because the pipeline is automated, REC training sets of arbitrary size can be produced without human collection or labeling; Harlequin already contains more than a million queries.
  • Since bounding boxes and non-target expressions are inherited from Flickr30k Entities, the synthetic data keeps spatial semantics such as perspective and object relations while adding color diversity and fresh backgrounds.
  • The rendering step can repair human annotation errors by generating images in which the referred object actually appears inside its stated box.
  • Even unnatural combinations such as a blue dog yield usable supervision, indicating that models can treat color as a disambiguating cue rather than relying on memorized color-object co-occurrences.

Reading between the lines

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

  • Editorial inference: the same color-swap setup could be used as a controlled diagnostic for whether a REC model leans on color versus object identity, since one scene can be re-rendered many times with only the expression's color changed.
  • Editorial inference: extending the variation function to attributes that change boxes, such as size or location, could convert this pipeline from appearance-only augmentation into a generator of compositional-reasoning benchmarks.
  • Editorial inference: a direct test of whether the gains come from accurate synthetic supervision or simply from more data would be to pre-train on Harlequin queries paired with randomly shuffled images and compare against the faithful-image setting.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 4 minor

Summary. The paper introduces Harlequin, a synthetic dataset for Referring Expression Comprehension (REC) generated from Flickr30k Entities by replacing color adjectives in referring expressions and captions with alternatives sampled from a 12-color vocabulary, then rendering new images with GLIGEN conditioned on the modified text and the original bounding-box layout. The authors pre-train three REC models (TransVG, VLTVG, LGR-NET) on Harlequin and fine-tune them on RefCOCO, RefCOCO+, and RefCOCOg, reporting accuracy gains over from-scratch training in most splits and larger gains on test subsets whose referring expressions contain a color attribute. The paper argues that this demonstrates that fully synthetic, annotation-free data can serve as effective pre-training for REC and that the generated data can correct some human annotation errors.

Significance. If the central claim is corroborated, the work is significant: it offers a scalable, low-cost route to REC training data, a reusable recipe based on color substitution and grounded text-to-image generation, and a concrete dataset (286k images / 1.09M queries) with released code. The paper deserves credit for evaluating three different architectures across three benchmark families and for including a color-subset analysis. However, the paper currently does not measure whether GLIGEN actually renders the substituted color on the referred object inside the provided box, and it lacks both a real-data pre-training control and variance/significance reporting. Those gaps are load-bearing for the attribution of the observed gains to the proposed synthetic generation mechanism.

major comments (3)
  1. [Section 4.2 and Section 5, Fig. 3] The pipeline's stated mechanism is color-driven generation, and the quality of the supervision signal depends on GLIGEN faithfully rendering the substituted color on the referred object and placing that object inside the given bounding box. The only evidence offered is the qualitative examples in Fig. 3 and the assertion in Section 5 that generation 'fixes some errors in the human-annotated labels.' The paper should report a quantitative fidelity measurement on a random sample of Harlequin, for example the fraction of generated images in which (a) the referred object's box has at least a threshold IoU with the annotation and (b) the replaced color is recognized by a color classifier or human judgment. Without such a measurement, the 1M+ annotations may be systematically misaligned with the pixels, and the Table 1 gains could be explained by pre-training on a large but noisy corpus rather than by accurate color-conditioned synthetic supervision.
  2. [Section 6.3, Table 1] Table 1 compares Synth→Real fine-tuning only against a from-scratch baseline (Real). It does not compare against pre-training on real data, such as Flickr30k Entities, from which Harlequin is derived. Because Harlequin contains about four times as many referring expressions as Flickr30k Entities and the authors pre-train for 60 epochs, the observed improvements (for example +4.11 for LGR-NET on RefCOCOg test) could stem from additional pre-training data or iterations rather than from the synthetic color-driven generation itself. A controlled baseline that pre-trains on Flickr30k Entities with the same schedule, and ideally a matched amount of data, is necessary to attribute the gains to the proposed generation pipeline.
  3. [Section 6.3, Table 1, and Section 6.4, Table 2] The paper reports single-run accuracy values, yet several differences between Real and Synth→Real are small (e.g., TransVG −0.10 on RefCOCO+ testB, VLTVG −0.53 on RefCOCOg val, LGR-NET +0.29 on RefCOCO+ val). With no error bars, no repeated-seed experiments, and no significance testing, the conclusion that 'pre-training on artificial data is beneficial' is not yet statistically supported. At minimum, the authors should report mean and standard deviation over at least three seeds per condition and state whether the gains are consistent across seeds.
minor comments (4)
  1. [Abstract and Section 4] The abstract says the approach 'eliminates manual data collection and annotation,' but the pipeline uses Flickr30k Entities human annotations as seeds for the Annotation Generation Engine. The claim should be qualified to clarify that it eliminates manual annotation of the newly generated synthetic images while still relying on seed annotations from an existing dataset.
  2. [Section 5 and Conclusion] The paper repeatedly calls Harlequin 'the first totally synthetic generated dataset' for REC. This claim is hard to verify and is not load-bearing for the main contribution. It would be safer to state that Harlequin is, to the authors' knowledge, the first fully synthetic REC dataset of this kind, or to cite any prior synthetic REC datasets explicitly.
  3. [Section 6.1] The implementation-details paragraph mixes pre-training, fine-tuning, and baseline training schedules in one long sentence. Separating the three settings into distinct descriptions would improve readability and reproducibility.
  4. [Section 6.4] The sentence 'These results were expected' and the subsequent assertion that the improvement is 'solely guided by learning patterns from data' are not established by the experiments. The authors should either provide an analysis supporting this mechanism or soften the claim.

Circularity Check

0 steps flagged · score 2.0 of 10

No significant circularity: the synthetic-data pre-training claim is evaluated on external human-annotated benchmarks, and the lone self-citation is not load-bearing.

full rationale

Harlequin is generated from Flickr30k Entities annotations by the AGE color-replacement function (Eqs. 4-5) and rendered by GLIGEN; no parameter is fitted to RefCOCO/RefCOCO+/RefCOCOg, and the fine-tuning evaluation uses standard accuracy at IoU>=0.5 (Section 6.2). The claimed improvements in Tables 1-2 are therefore measured against data external to the generation pipeline. The only self-citation is ref. 28 in Section 2 (weakly-supervised grounding), where it is cited as an example of limited-supervision work, not as evidence for Harlequin's validity. Section 6.4's explanation that color-only test subsets improve because Harlequin is color-centric is a post-hoc interpretation, not a reduction of the result to the input. The paper does assert without quantitative verification that GLIGEN renders the replaced color and keeps objects inside the provided boxes (Section 4.2 and Section 5), which is a correctness risk, but that unmeasured fidelity is not a circularity because the benchmark evaluation is still external and no fitted constant is recycled as a prediction.

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

The central claim rests on three domain assumptions about the generative model and the seed data, plus three design choices that are not fitted to the target benchmarks. No new conceptual entities are introduced.

free parameters (3)
  • color vocabulary = 12 colors: black, gray, white, red, orange, yellow, green, cyan, blue, purple, pink, brown
    Chosen from RSVG [44] to vary referring expressions. The set and its size determine the diversity of generated annotations and images.
  • number of color variations per referring expression = 6 (= |C|/2)
    Chosen as a trade-off between dataset size and variability; directly controls the scale of Harlequin (1.09M queries).
  • seed sample selection = only Flickr30k Entities samples whose referring expressions contain a color attribute
    The pipeline only processes seeds with color attributes; this filtering defines Harlequin's domain and limits generalization to color-bearing expressions.
assumptions (3)
  • domain assumption GLIGEN, a frozen pre-trained text-to-image model, can generate images consistent with caption and per-object region descriptions and boxes.
    Section 4.2 implements the Image Generation Engine with frozen GLIGEN; if this assumption fails, the generated images do not match the annotations and the dataset's supervision is corrupted.
  • domain assumption Replacing the color attribute in a referring expression does not change the object's location, so the original bounding box remains valid in the newly generated image.
    Sections 4 and 4.1 keep all locations fixed while varying color; the generated image is expected to place the recolored object in the same box. This is not verified quantitatively.
  • domain assumption Flickr30k Entities annotations are a reliable enough starting point, with correct bounding boxes and consistent referring expressions, for generating seeds.
    Section 3 and 5 use Flickr30k Entities as seeds; the paper notes some label errors but does not quantify how often they occur or whether they propagate.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Harlequin: Color-driven Generation of Synthetic Data for Referring Expression Comprehension." pith.science (2026). https://pith.science/paper/AM5LX225

@misc{pith2026241114807,
  author       = {Pith},
  title        = {Pith review of: Harlequin: Color-driven Generation of Synthetic Data for Referring Expression Comprehension},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/AM5LX225}},
  note         = {Machine review of arXiv:2411.14807}
}
read the original abstract

Referring Expression Comprehension (REC) aims to identify a particular object in a scene by a natural language expression, and is an important topic in visual language understanding. State-of-the-art methods for this task are based on deep learning, which generally requires expensive and manually labeled annotations. Some works tackle the problem with limited-supervision learning or relying on Large Vision and Language Models. However, the development of techniques to synthesize labeled data is overlooked. In this paper, we propose a novel framework that generates artificial data for the REC task, taking into account both textual and visual modalities. At first, our pipeline processes existing data to create variations in the annotations. Then, it generates an image using altered annotations as guidance. The result of this pipeline is a new dataset, called Harlequin, made by more than 1M queries. This approach eliminates manual data collection and annotation, enabling scalability and facilitating arbitrary complexity. We pre-train three REC models on Harlequin, then fine-tuned and evaluated on human-annotated datasets. Our experiments show that the pre-training on artificial data is beneficial for performance.

Figures

Figures reproduced from arXiv: 2411.14807 by the authors.

Figure 1
Figure 1. Annotations required by the Re￾ferring Expression Comprehension task. In this example, the image has one caption with three referring expressions. Each refer￾ring expression is accompanied by the loca￾tion of the referred object (bounding box). These datasets are human-labeled and consist of triplets composed of an image, a referring expression, and a bounding box [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Our pipeline. It processes existing samples from Flickr30k Entities data. We se￾lect the ones characterized by at least one color attribute in their referring expressions. The Annotation Generation Engine processes the sample’s caption, referring expres￾sions and locations where the color attribute is replaced with a randomly chosen color. The caption is updated accordingly. Then, the Image Generation Engine creates… view at source ↗
Figure 3
Figure 3. Examples produced by our pipeline. On the left, we show reference images along with their annotations from Flickr30k Entities. On the right, we report some generated variations. Colors are altered and guide, along with objects’ locations, the image synthesis. 5 Harlequin Dataset We introduce the first totally synthetic generated dataset for the Referring Ex￾pression Comprehension task, termed Harlequin,2 collected v… view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Dataset statistics. We report the number of images and referring expressions per dataset on the left and right, respectively. Harlequin is highlighted in orange. Harlequin follows Flickr30k Entities’ data splits. It provides 988,342 annotations over 259,930 images for …

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

45 extracted references · 40 canonical work pages

  1. [1]

    CoRR abs/2304.00500 (2023) 3

    Amoroso, R., Morelli, D., Cornia, M., Baraldi, L., Bimbo, A.D., Cucchiara, R.: Parents and children: Distinguishing multimodal deepfakes from natural images. CoRR abs/2304.00500 (2023) 3

    arXiv 2023

  2. [2]

    In: Proc

    Arora, S., Narayan, A., Chen, M.F., Orr, L.J., Guha, N., Bhatia, K., Chami, I., Ré, C.: Ask me anything: A simple strategy for prompting language models. In: Proc. of the International Conference on Learning Representations (ICLR) (2023) 13

    work page 2023

  3. [3]

    Blender Online Community: Blender - a 3D modelling and rendering package (2024), http://www.blender.org 3

    work page 2024

  4. [4]

    In: Proc

    Carion, N., Massa, F., Synnaeve, G., Usunier, N., Kirillov, A., Zagoruyko, S.: End- to-end object detection with transformers. In: Proc. of the European Conference on Computer Vision (ECCV) (2020) 10

    work page 2020

  5. [5]

    In: Proc

    Chen, Z., Wang, P., Ma, L., Wong, K.K., Wu, Q.: Cops-ref: A new dataset and task on compositional referring expression comprehension. In: Proc. of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2020) 2

    work page 2020

  6. [6]

    In: Proc

    Danelljan, M., Shahbaz Khan, F., Felsberg, M., Van de Weijer, J.: Adaptive color attributes for real-time visual tracking. In: Proc. of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2014) 6

    work page 2014

  7. [7]

    In: Proc

    Deng, J., Yang, Z., Chen, T., Zhou, W., Li, H.: Transvg: End-to-end visual ground- ing with transformers. In: Proc. of the IEEE/CVF International Conference on Computer Vision (ICCV) (2021) 1, 3, 10, 11, 12

    work page 2021

  8. [8]

    In: Proc

    Devlin, J., Chang, M., Lee, K., Toutanova, K.: BERT: pre-training of deep bidi- rectional transformers for language understanding. In: Proc. of the Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL-HLT) (2019) 10

    work page 2019

Show all 45 references
  1. [9]

    In: Proc

    Dosovitskiy, A., Ros, G., Codevilla, F., López, A.M., Koltun, V.: CARLA: an open urban driving simulator. In: Proc. of the Conference on Robot Learning (CoRL) (2017) 3

    2017

  2. [10]

    Escalante, H.J., Hernández, C.A., González, J.A., López-López, A., Montes-y- Gómez, M., Morales, E.F., Sucar, L.E., Pineda, L.V., Grubinger, M.: The seg- mented and annotated IAPR TC-12 benchmark. Comput. Vis. Image Underst. 114(4), 419–428 (2010) 4

    2010

  3. [11]

    CoRRabs/2305.15393 (2023) 13

    Feng, W., Zhu, W., Fu, T., Jampani, V., Akula, A.R., He, X., Basu, S., Wang, X.E., Wang, W.Y.: Layoutgpt: Compositional visual planning and generation with large language models. CoRRabs/2305.15393 (2023) 13

    2023 arXiv

  4. [12]

    CoRRabs/1605.06457 (2016) 3, 4

    Gaidon, A., Wang, Q., Cabon, Y., Vig, E.: Virtual worlds as proxy for multi-object tracking analysis. CoRRabs/1605.06457 (2016) 3, 4

    2016 arXiv

  5. [13]

    of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2016) 10

    He,K.,Zhang,X.,Ren,S.,Sun,J.:Deepresiduallearningforimagerecognition.In: Proc. of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2016) 10

    2016

  6. [14]

    In: Proc

    Jiang, H., Lin, Y., Han, D., Song, S., Huang, G.: Pseudo-q: Generating pseudo language queries for visual grounding. In: Proc. of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2022) 2

    2022

  7. [15]

    In: Proc

    Jin, J., Ye, J., Lin, X., He, L.: Pseudo-query generation for semi-supervised vi- sual grounding with knowledge distillation. In: Proc. of the IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) (2023) 2

    2023

  8. [16]

    CoRRabs/1809.02627 (2018) 3 Harlequin: Color-driven Generation of Synthetic Data for REC 15

    Juliani, A., Berges, V., Vckay, E., Gao, Y., Henry, H., Mattar, M., Lange, D.: Unity: A general platform for intelligent agents. CoRRabs/1809.02627 (2018) 3 Harlequin: Color-driven Generation of Synthetic Data for REC 15

    2018 arXiv

  9. [17]

    In: Proc

    Kamath, A., Singh, M., LeCun, Y., Synnaeve, G., Misra, I., Carion, N.: MDETR - modulated detection for end-to-end multi-modal understanding. In: Proc. of the IEEE/CVF International Conference on Computer Vision (ICCV) (2021) 2

    2021

  10. [18]

    In: Proc

    Kazemzadeh, S., Ordonez, V., Matten, M., Berg, T.L.: Referitgame: Referring to objects in photographs of natural scenes. In: Proc. of the Conference on Empirical Methods in Natural Language Processing (EMNLP) (2014) 1, 6

    2014

  11. [19]

    In: Proc

    Le,H.,Mensink,T.,Das,P.,Karaoglu,S.,Gevers,T.:EDEN:multimodalsynthetic dataset of enclosed garden scenes. In: Proc. of the IEEE/CVF Winter Conference on Applications of Computer Vision (WACV) (2021) 3

    2021

  12. [20]

    In: Proc

    Li, M., Sigal, L.: Referring transformer: A one-step approach to multi-task vi- sual grounding. In: Proc. of Advances in Neural Information Processing Systems (NeurIPS) (2021) 1

    2021

  13. [21]

    In: Proc

    Li, Y., Liu, H., Wu, Q., Mu, F., Yang, J., Gao, J., Li, C., Lee, Y.J.: GLIGEN: open- set grounded text-to-image generation. In: Proc. of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2023) 2, 4, 5, 7

    2023

  14. [22]

    In: Proc

    Lin, T., Maire, M., Belongie, S.J., Hays, J., Perona, P., Ramanan, D., Dollár, P., Zitnick, C.L.: Microsoft COCO: common objects in context. In: Proc. of the European Conference on Computer Vision (ECCV) (2014) 4

    2014

  15. [23]

    In: Proc

    Liu, Z., Lin, Y., Cao, Y., Hu, H., Wei, Y., Zhang, Z., Lin, S., Guo, B.: Swin transformer: Hierarchical vision transformer using shifted windows. In: Proc. of the IEEE/CVF International Conference on Computer Vision (ICCV) (2021) 10

    2021

  16. [24]

    In: Proc

    Lu, J., Rao, J., Chen, K., Guo, X., Zhang, Y., Sun, B., Yang, C., Yang, J.: Evalu- ation and enhancement of semantic grounding in large vision-language models. In: Proc. of the AAAI Workshop on Responsible Language Models (ReLM) (2024) 6

    2024

  17. [25]

    IEEE Transactions on Circuits and Systems for Video Technology (TCSVT)in press (2024) 3, 10, 11

    Lu, M., Li, R., Feng, F., Ma, Z., Wang, X.: LGR-NET: Language Guided Reasoning Network for Referring Expression Comprehension. IEEE Transactions on Circuits and Systems for Video Technology (TCSVT)in press (2024) 3, 10, 11

    2024

  18. [26]

    In: Proc

    Mao, J., Huang, J., Toshev, A., Camburu, O., Yuille, A.L., Murphy, K.: Gen- eration and comprehension of unambiguous object descriptions. In: Proc. of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2016) 1

    2016

  19. [27]

    In: Proc

    Plummer, B.A., Wang, L., Cervantes, C.M., Caicedo, J.C., Hockenmaier, J., Lazeb- nik, S.: Flickr30k entities: Collecting region-to-phrase correspondences for richer image-to-sentence models. In: Proc. of the IEEE/CVF International Conference on Computer Vision (ICCV) (2015) 1

    2015

  20. [28]

    In: Proc

    Rigoni, D., Parolari, L., Serafini, L., Sperduti, A., Ballan, L.: Weakly-supervised visual-textual grounding with semantic prior refinement. In: Proc. of the British Machine Vision Conference (BMVC) (2023) 2

    2023

  21. [29]

    In: Proc

    Rohrbach, A., Rohrbach, M., Hu, R., Darrell, T., Schiele, B.: Grounding of textual phrases in images by reconstruction. In: Proc. of the European Conference on Computer Vision (ECCV) (2016) 1

    2016

  22. [30]

    In: Proc

    Rombach, R., Blattmann, A., Lorenz, D., Esser, P., Ommer, B.: High-resolution image synthesis with latent diffusion models. In: Proc. of the IEEE/CVF Confer- ence on Computer Vision and Pattern Recognition (CVPR) (2022) 5, 7

    2022

  23. [31]

    CoRR abs/2208.04052 (2022) 3

    Roovere, P.D., Moonen, S., Michiels, N., Wyffels, F.: Dataset of industrial metal objects. CoRR abs/2208.04052 (2022) 3

    2022 arXiv

  24. [32]

    In: Proc

    Su, W., Miao, P., Dou, H., Wang, G., Qiao, L., Li, Z., Li, X.: Language adaptive weight generation for multi-task visual grounding. In: Proc. of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2023) 3 16 Luca Parolari, Elena Izzo, and Lamberto Ballan

    2023

  25. [33]

    In: Proc

    Tanaka, M., Itamochi, T., Narioka, K., Sato, I., Ushiku, Y., Harada, T.: Generating easy-to-understand referring expressions for target identifications. In: Proc. of the IEEE/CVF International Conference on Computer Vision (ICCV) (2019) 2

    2019

  26. [34]

    In: Proc

    Wang, J., Specia, L.: Phrase localization without paired training examples. In: Proc. of the IEEE/CVF International Conference on Computer Vision (ICCV) (2019) 1, 2

    2019

  27. [35]

    IEEE Transactions on Multimedia (TMM) 25, 9330–9342 (2023) 3

    Wang,W.,Pagnucco,M.,Xu,C.,Song,Y.:InterREC:AnInterpretableMethodfor Referring Expression Comprehension. IEEE Transactions on Multimedia (TMM) 25, 9330–9342 (2023) 3

    2023

  28. [36]

    In: Proc

    Yang, L., Xu, Y., Yuan, C., Liu, W., Li, B., Hu, W.: Improving visual ground- ing with visual-linguistic verification and iterative reasoning. In: Proc. of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2022) 3, 10, 11, 12, 13

    2022

  29. [37]

    In: Proc

    Yang, Z., Gan, Z., Wang, J., Hu, X., Ahmed, F., Liu, Z., Lu, Y., Wang, L.: Unitab: Unifying text and box outputs for grounded vision-language modeling. In: Proc. of the European Conference on Computer Vision (ECCV) (2022) 2

    2022

  30. [38]

    In: Proc

    Yang, Z., Gong, B., Wang, L., Huang, W., Yu, D., Luo, J.: A fast and accurate one-stage approach to visual grounding. In: Proc. of the IEEE/CVF International Conference on Computer Vision (ICCV) (2019) 3

    2019

  31. [39]

    In: Proc

    Yang, Z., Wang, J., Gan, Z., Li, L., Lin, K., Wu, C., Duan, N., Liu, Z., Liu, C., Zeng, M., Wang, L.: Reco: Region-controlled text-to-image generation. In: Proc. of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2023) 4

    2023

  32. [40]

    In: Proc

    Ye, J., Tian, J., Yan, M., Yang, X., Wang, X., Zhang, J., He, L., Lin, X.: Shifting more attention to visual backbone: Query-modulated refinement networks for end- to-end visual grounding. In: Proc. of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2022) 3

    2022

  33. [41]

    Transactions of the Association for Computational Linguistics (TACL)2, 67–78 (2014) 4

    Young, P., Lai, A., Hodosh, M., Hockenmaier, J.: From image descriptions to visual denotations: New similarity metrics for semantic inference over event descriptions. Transactions of the Association for Computational Linguistics (TACL)2, 67–78 (2014) 4

    2014

  34. [42]

    In: Proc

    Yu, L., Lin, Z., Shen, X., Yang, J., Lu, X., Bansal, M., Berg, T.L.: Mattnet: Mod- ular attention network for referring expression comprehension. In: Proc. of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2018) 1, 3

    2018

  35. [43]

    In: Proc

    Yu, L., Poirson, P., Yang, S., Berg, A.C., Berg, T.L.: Modeling context in referring expressions. In: Proc. of the European Conference on Computer Vision (ECCV) (2016) 1

    2016

  36. [44]

    IEEE Transactions on Geoscience and Remote Sensing 61, 1–13 (2023) 7

    Zhan, Y., Xiong, Z., Yuan, Y.: RSVG: exploring data and models for visual ground- ing on remote sensing data. IEEE Transactions on Geoscience and Remote Sensing 61, 1–13 (2023) 7

    2023

  37. [45]

    In: Proc

    Zhang, L., Rao, A., Agrawala, M.: Adding conditional control to text-to-image dif- fusion models. In: Proc. of the IEEE/CVF International Conference on Computer Vision (ICCV) (2023) 4

    2023

Pith tools

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