Pith. sign in

REVIEW 3 major objections 6 minor 27 references

Learning from Multimodal Pseudo-Labels for Robust Open-Vocabulary Instance and Panoptic Segmentation

T0 review · 3 major / 6 minor · reviewed 2026-08-16 · deepseek-v4-flash

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

Pith's one-line read The paper claims that adding CLIP-filtered synonyms, a semantic-consistency loss, and a masked-caption-reconstruction loss lifts novel-class mask average precision by 6.5 and 6.8 points over the same pseudo-mask supervision.

desk verdict A competent, incremental pseudo-label pipeline for open-vocabulary segmentation whose headline gain over its own control is not yet isolated: the control (CGG†) may have a different text encoder than the proposed method, so the +6.5/+6.8 AP claim could be a BERT-to-CLIP effect. read the letter →

arxiv 2608.11681 v1 pith:733PCQSK submitted 2026-08-12 cs.CV cs.AIcs.MM

Duy Tran Thanh , Yeejin Lee , Byeongkeun Kang This is my paper · ORCID
classification cs.CVcs.AIcs.MM
keywords open-vocabularyinstancesegmentationopen-setpanopticmultimodalpseudo-labelingCLIP-guidedsynonymfilteringsemanticconsistencylosscaptionreconstructiontarget-vocabulary-assistedprotocolvision-languagealignment
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

This paper tries to show that open-vocabulary instance and panoptic segmentation can be improved without extra manual annotations by supervising the model with multimodal pseudo-labels: pseudo masks, pseudo captions, and visually verified synonyms generated by pre-trained vision-language models. On the COCO benchmark under its target-vocabulary-assisted protocol, the method reports 51.6 novel-class AP in the constrained setting and 50.4 in the generalized setting, compared with 45.1 and 43.6 for CGG retrained with the same pseudo masks. The three added objectives, extended grounding with synonyms, semantic consistency, and masked caption reconstruction, account for the gap. If this is right, language-side supervision derived from captions and synonyms adds real segmentation ability on top of better pseudo-mask quality, at about 2.4% extra inference FLOPs.

What carries the argument

The core mechanism is joint training supervision built from three automatically generated signal types and three losses. Pseudo masks from a text-prompted segmenter give novel-class visual supervision; pseudo captions from a vision-language assistant add lexical diversity; CLIP text-image similarity with a top-1 ranking and a threshold of 0.4 filters candidate synonyms to those actually consistent with the masked region. The losses are an extended grounding loss $L_{gr}$ that matches region features to both novel-class names and synonyms, a semantic consistency loss that penalizes $(f_i^T(t_{nov} - t_{syn}))^2$ to pull embeddings of synonymous words together, and a generative caption reconstruction loss $L_{recon}$ that reconstructs masked caption tokens conditioned on image features. Together they make the model treat a class name and its paraphrases as the same visual concept. All auxiliary generators are removed at inference, leaving the segmenter plus CLIP text embeddings.

What would settle it

Train the proposed model with pseudo-masks generated from a vocabulary disjoint from the test classes, such as generic descriptions or a different held-out split, and evaluate on the original novel classes; if novel-class AP falls back to the CGG† level, the reported gains come from target-vocabulary assistance rather than from the three proposed losses.

Watch

Extended reading notes

Core claim

The paper's central claim is that a multimodal pseudo-labeling pipeline together with three complementary training losses improves generalization to novel classes beyond what higher-quality pseudo masks alone provide. The pipeline uses a text-prompted segmenter to create pseudo masks for novel classes, a vision-language assistant to write descriptive pseudo captions, and CLIP similarity to keep only visually grounded synonyms. During training, the segmenter is pushed to align region features with both class names and synonyms, to keep synonym and class embeddings consistent, and to reconstruct masked caption tokens from visual features. In the controlled comparison against CGG retrained with the same pseudo masks, novel-class AP rises from 45.1 to 51.6 (constrained) and from 43.6 to 50.4 (generalized); in open-set panoptic segmentation, unknown-class PQ improves by 18.0, 11.5, and 7.3 points under the 20%, 10%, and 5% unknown settings, with a small drop on known classes.

Load-bearing premise

The evaluation assumes it is acceptable to tell the pseudo-mask generator the exact names of the classes that will be tested, so the reported gains reflect help from the target vocabulary rather than purely never-seen-category generalization.

Editorial extensions

If this is right

  • Training with CLIP-verified synonyms should make the segmenter more tolerant of paraphrases at test time, reducing failures when a class is named in words other than the training label.
  • Because the auxiliary generators are removed at inference, the added capability costs only about 2.4% more GFLOPs and 7.9% more parameters, so the approach can be used where inference efficiency matters.
  • The method shifts accuracy toward novel and unknown classes at a small cost to known classes in panoptic settings, so users can expect better coverage of rare categories in exchange for slightly lower base-class PQ.
  • Removing CLIP-based synonym filtering drops constrained novel AP from 51.6 to 47.1, indicating that visual verification of candidate words is a load-bearing part of the pipeline.

Reading between the lines

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

  • A fair test of genuine open-vocabulary generalization would generate pseudo-masks with a vocabulary disjoint from the test classes; the paper's protocol supplies the exact test class names, so the absolute AP numbers partly credit that assistance.
  • The CLIP-filtering design is portable: any candidate word list could be validated against masked image regions the same way, so the technique could transfer to other weakly supervised segmentation pipelines.
  • The small known-class drop in panoptic settings suggests the method trades some base-class precision for unknown-class coverage, a trade that may be worth making in applications that value rare-object detection more than base-class exactness.
  • Because the caption reconstruction loss depends on masked-token prediction, richer or multi-view pseudo captions could plausibly increase the lexical diversity the model is exposed to, possibly enlarging the gain beyond what this paper reports.
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

3 major / 6 minor

Summary. The paper proposes MCCF, a Mask2Former-based framework for open-vocabulary instance segmentation (OVIS) and open-set panoptic segmentation (OSPS). It automatically generates pseudo segmentation masks with Grounded SAM using the target novel-class vocabulary as prompts, pseudo captions with LLaVA, and visually grounded synonyms with CLIP-based filtering. During training it adds three objectives to the CGG baseline: an extended grounding loss, a semantic consistency loss, and a GPT-based caption reconstruction loss. On COCO, the method reports large gains over CGG and a controlled baseline CGG† in OVIS, and gains over CGG in OSPS.

Significance. If the reported controlled comparison is valid, the paper provides evidence that synonym-aware grounding and caption reconstruction can add meaningful improvements beyond stronger pseudo-mask quality. The introduction of CGG† as a control is a good design choice, and the cumulative ablation in Table 4 is internally consistent. The CLIP-filtering ablation in Table 5 also clearly demonstrates the value of visual grounding for synonym selection. However, the target-vocabulary-assisted protocol means the experiments do not measure open-vocabulary generalization in the usual zero-shot sense, and the CGG† baseline is not fully specified. The paper is honest about the protocol in Section 4.1, but the framing of the contributions needs to be adjusted to match what is actually measured.

major comments (3)
  1. [Section 4.1 (CGG†) and Section 3.2] The text encoder used by CGG† is not specified. Section 3.2 states that CGG uses BERT while MCCF uses the CLIP text encoder, and the classification, grounding, and consistency losses (Eqs. 2, 4, 5) all operate on CLIP text embeddings. If CGG† retains BERT, the headline +6.5/+6.8 AP improvements in Table 2 and the ablation increments in Table 4 conflate the text-encoder change with the proposed losses. The authors must state explicitly which text encoder CGG† uses; if it is BERT, an additional control using the CLIP text encoder with the original CGG losses is required.
  2. [Section 3.3.1 and Section 4.1 ('Pseudo-label vocabulary')] The target novel-class names are used as text prompts for Grounded SAM to generate pseudo-masks, so the exact test vocabulary is injected into the training pipeline. The evaluation therefore does not assess generalization to categories never seen during training, and the abstract's claims about 'unseen object categories' are not supported by the experiments. The paper should reformulate the claims in terms of a target-vocabulary-assisted protocol, and ideally report an additional experiment in which pseudo-masks are generated without the target-vocabulary prompts (e.g., from captions only) to quantify the resulting drop in novel-class AP.
  3. [Table 3 (OSPS)] The open-set panoptic segmentation comparison is not controlled like the OVIS comparison. CGG in Table 3 is the original method without the Grounded-SAM pseudo-mask annotations, so the unknown-class PQ gains (18.0, 11.5, and 7.3 points under the 20%, 10%, and 5% settings) conflate the effect of the pseudo-masks with the effect of the proposed losses. A CGG†-style baseline using the same pseudo-mask annotations should be reported for the OSPS settings to support the claim that the proposed objectives are responsible for the improvement.
minor comments (6)
  1. [Section 3.3.2, Eq. (5)] In Eq. (5), the notation f_i^T (t_i^nov - t_i^syn) is unclear about the dimension of f_i; if f_i is the decoder embedding for instance i, state this and clarify whether the loss is averaged over instances.
  2. [Section 4.1 (Implementation Details)] The relationship between the GPT-2 tokenizer/embedding layer and the CLIP text encoder should be clarified: the classification and grounding losses use CLIP text embeddings, while the caption reconstruction loss appears to use GPT-2 embeddings; state how these two text representations are reconciled.
  3. [Table 3] The column headers in Table 3 are confusing: 'PQTh' appears twice, and the superscripts 'Th' and 'St' are not consistently defined in the caption; the asterisk footnote is also difficult to parse.
  4. [Section 3.3.1 (Synonym filtering)] The average number of synonyms retained per class and the sensitivity of the results to the threshold τ=0.4 are not reported; this would help assess the stability of the filtering step.
  5. [Title] The title contains a stray space in 'Open-V ocabulary'; this should be corrected.
  6. [References] The reference for Loshchilov (2019) is incomplete; it should list all authors and the full proceedings.

Circularity Check

0 steps flagged · score 0.0 of 10

No circular reduction found; the target-vocabulary protocol and CGG† text-encoder ambiguity are validity concerns, not circularity.

full rationale

The paper's derivation chain is not circular. Pseudo-masks come from frozen Grounded SAM, captions from LLaVA, and synonyms from CLIP-filtered LLaVA output; the three training losses (Eqs. 4–6) are additional objectives trained on those fixed pseudo-labels, and the controlled baseline CGG† is trained with the same Grounded-SAM masks under the original CGG losses. The reported +6.5/+6.8 novel AP over CGG† compares methods that differ by the proposed losses, so the headline claim has independent empirical content. The target-vocabulary-assisted protocol does inject novel class names into pseudo-label generation, but the paper explicitly discloses this and does not claim a category-agnostic protocol; this is an external-validity limitation rather than an equation-level circularity. Similarly, Section 3.2's BERT-versus-CLIP text-encoder difference could confound the CGG† comparison if CGG† retains BERT, but that is an experimental-control ambiguity, not a reduction of the claimed result to its inputs. The appended Limitations section reports only computational-scale constraints and does not assert any circular dependency. No self-definitional step, fitted-parameter-renamed-as-prediction, or load-bearing self-citation chain is present.

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

The method assumes the outputs of three pre-trained models (Grounded SAM, LLaVA, CLIP) are fit to serve as training targets, and fixes two threshold and weight hyperparameters by hand. No new objects, forces, or mathematical structures are proposed.

free parameters (3)
  • CLIP synonym filtering threshold τ = 0.4
    Empirically selected in Section 3.3.1 to balance retaining visually consistent synonyms versus suppressing noisy words; no sensitivity analysis is provided.
  • Loss weights λ_cls, λ_mask, λ_gr, λ_recon = 2, 5, 2, 2
    Set in Section 3.3.2 with no ablation; the balance affects the relative contribution of the new losses.
  • Masked token ratio for caption reconstruction = not specified
    Section 3.3.2 says 'a subset of tokens is randomly replaced' with no specified ratio, making the objective's difficulty a hidden free parameter.
assumptions (4)
  • domain assumption Grounded SAM produces accurate pseudo-masks when prompted with novel class names.
    Section 3.3.1 uses Grounding DINO and SAM outputs as training masks; if these pseudo-masks are noisy, the model learns from noisy labels.
  • domain assumption CLIP similarity between masked image regions and candidate words measures visual grounding.
    Used in Section 3.3.1 (Eq. 1) to filter synonyms; assumes CLIP's global image-text alignment extends to masked, cropped regions.
  • domain assumption LLaVA generates diverse, semantically correct pseudo-captions from the class-label prompt.
    Pseudo-captions are the source of candidate synonyms; LLaVA hallucination would propagate errors.
  • standard math Mask2Former, CLIP, and the grounding loss formulation from prior work are valid background.
    Baseline architecture and grounding loss are adopted from Cheng et al. 2022 and Wu et al. 2023.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Learning from Multimodal Pseudo-Labels for Robust Open-Vocabulary Instance and Panoptic Segmentation." pith.science (2026). https://pith.science/paper/733PCQSK

@misc{pith2026260811681,
  author       = {Pith},
  title        = {Pith review of: Learning from Multimodal Pseudo-Labels for Robust Open-Vocabulary Instance and Panoptic Segmentation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/733PCQSK}},
  note         = {Machine review of arXiv:2608.11681}
}
read the original abstract

This work addresses the challenge of open-vocabulary instance segmentation (OVIS) and open-set panoptic segmentation (OSPS), which aim to recognize both predefined and unseen object categories without exhaustive human annotations. Existing methods often suffer from noisy pseudo-masks, limited visual-textual grounding, and difficulty handling synonyms or out-of-vocabulary (OOV) words. To overcome these challenges, we propose a multimodal framework that leverages pre-trained vision-language models for automatic pseudo-label generation, CLIP-guided synonym filtering, and GPT-based caption reconstruction. In our target-vocabulary-assisted pseudo-labeling setting, the framework first constructs pseudo segmentation masks, descriptive captions, and semantically aligned synonym sets using Grounded SAM, LLaVA, and CLIP, providing multimodal supervision without manual annotation. We then enhance visual-textual alignment through three complementary training objectives: an extended grounding loss that incorporates visually grounded synonyms, a semantic consistency loss, and a generative caption reconstruction loss. Extensive experiments on the COCO dataset demonstrate that the proposed method consistently outperforms previous state-of-the-art approaches under this protocol, achieving substantial improvements on both OVIS and OSPS benchmarks.

Figures

Figures reproduced from arXiv: 2608.11681 by the authors.

Figure 1
Figure 1. Overall framework of MCCF. The proposed framework consists of three stages: multimodal pseudo-label construction, training with multimodal su [PITH_FULL_IMAGE:figures/full_fig_p005_1.png] view at source ↗
Figure 2
Figure 2. Multimodal pseudo-label construction. Grounding DINO and SAM [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Training framework of MCCF. The Mask2Former-based segmenter is trained with pseudo masks, class names, pseudo synonyms, and captions. In addition [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: Qualitative comparison of open-vocabulary instance segmentation [PITH_FULL_IMAGE:figures/full_fig_p009_4.png]
Figure 5
Figure 5. Figure 5: Qualitative comparison of open-set panoptic segmentation results between the proposed method (second row) and the previous state-of-the-art method, [PITH_FULL_IMAGE:figures/full_fig_p010_5.png]
Figure 6
Figure 6. Figure 6: Additional qualitative results for OVIS and OSPS. The first and second rows correspond to OVIS and OSPS examples, respectively. The captions below [PITH_FULL_IMAGE:figures/full_fig_p010_6.png]
Figure 7
Figure 7. Figure 7: Failure cases in OVIS. The examples show typical errors caused by over-segmentation, cluttered scenes, partial occlusion, low illumination, and ambiguous [PITH_FULL_IMAGE:figures/full_fig_p011_7.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

27 extracted references · 20 canonical work pages

  1. [2]

    Neurocomputing 665, 132229

    Efficient redundancy reduction for open-vocabulary semantic segmentation. Neurocomputing 665, 132229. URL:https://www.sciencedirect. com/science/article/pii/S0925231225029017, doi:https://doi. org/10.1016/j.neucom.2025.132229. Cheng, B., Misra, I., Schwing, A.G., Kirillov, A., Girdhar, R.,

    work page arXiv 2025

  2. [7]

    (Eds.), Computer Vision – ECCV 2022, Springer Nature Switzerland, Cham

    Scaling open-vocabulary image segmentation with image-level labels, in: Avidan, S., Brostow, G., Cissé, M., Farinella, G.M., Hassner, T. (Eds.), Computer Vision – ECCV 2022, Springer Nature Switzerland, Cham. pp. 540–557. Gupta, A., Dollar, P., Girshick, R.,

    work page 2022

  3. [8]

    7010–7021

    Open- vocabulary instance segmentation via robust cross-modal pseudo-labeling, in: 2022 IEEE/CVF Conference on Computer Vision and Pattern Recogni- tion (CVPR), pp. 7010–7021. Hwang, J., Oh, S.W., Lee, J.Y ., Han, B.,

    work page 2022

  4. [9]

    Neurocomput- ing 659, 131790

    Ov-kfa: Open- vocabulary object detection via key feature alignment. Neurocomput- ing 659, 131790. URL:https://www.sciencedirect.com/science/ article/pii/S0925231225024622, doi:https://doi.org/10.1016/ j.neucom.2025.131790. Kim, D.M., Lee, S., Kang, B., 2025a. Completely weakly supervised class- incremental learning for semantic segmentation. Pattern Reco...

    work page arXiv 2025

  5. [12]

    Neurocomput- ing 651, 131018

    Marins3d: An open-vocabulary 3d instance segmentation model with mask refinement. Neurocomput- ing 651, 131018. URL:https://www.sciencedirect.com/science/ article/pii/S092523122501690X, doi:https://doi.org/10.1016/ j.neucom.2025.131018. Li, X., Yuan, H., Li, W., Ding, H., Wu, S., Zhang, W., Li, Y ., Chen, K., Loy, C.C.,

    work page arXiv

  6. [15]

    (Eds.), Advances in Neural Information Processing Sys- tems, Curran Associates, Inc

    Visual instruction tuning, in: Oh, A., Naumann, T., Globerson, A., Saenko, K., Hardt, M., Levine, S. (Eds.), Advances in Neural Information Processing Sys- tems, Curran Associates, Inc.. pp. 34892–34916. URL:https: //proceedings.neurips.cc/paper_files/paper/2023/file/ 6dcf277ea32ce3288914faf369fe6de0-Paper-Conference.pdf. Liu, S., Zeng, Z., Ren, T., Li, F...

    work page arXiv 2023

  7. [16]

    Neurocomputing 677, 133088

    Spsrl: Open-vocabulary semantic segmentation with spatial prior and semantic relation learning. Neurocomputing 677, 133088. URL:https://www.sciencedirect.com/science/ article/pii/S0925231226004856, doi:https://doi.org/10.1016/ j.neucom.2026.133088. Qin, J., Wu, J., Yan, P., Li, M., Yuxi, R., Xiao, X., Wang, Y ., Wang, R., Wen, S., Pan, X., Wang, X.,

    work page arXiv

  8. [18]

    URL:https://arxiv.org/abs/2401.14159, arXiv:2401.14159

    Grounded sam: Assembling open-world models for diverse visual tasks. URL:https://arxiv.org/abs/2401.14159, arXiv:2401.14159. Sennrich, R., Haddow, B., Birch, A.,

    arXiv

Show all 27 references
  1. [20]

    (Eds.), Advances in Neural Information Pro- cessing Systems, Curran Associates, Inc

    Learning equivariant segmentation with instance-unique querying, in: Koyejo, S., Mohamed, S., Agarwal, A., Belgrave, D., Cho, K., Oh, A. (Eds.), Advances in Neural Information Pro- cessing Systems, Curran Associates, Inc.. pp. 12826–12840. URL:https: //proceedings.neurips.cc/p...

    2022 doi

  2. [21]

    21881–21891

    Be- trayed by captions: Joint caption grounding and generation for open vocab- ulary instance segmentation, in: 2023 IEEE/CVF International Conference on Computer Vision (ICCV), pp. 21881–21891. Wu, J., Li, X., Xu, S., Yuan, H., Ding, H., Yang, Y ., Li, X., Zhang, J., Tong, Y ...

    2023

  3. [22]

    Neurocomput- ing 636, 129982

    Gcd-net: Global consciousness- driven open-vocabulary semantic segmentation network. Neurocomput- ing 636, 129982. URL:https://www.sciencedirect.com/science/ article/pii/S092523122500654X, doi:https://doi.org/10.1016/ j.neucom.2025.129982. Xu, H.M., Chen, H., Liu, L., Yin, Y .,

  4. [24]

    (Eds.), Computer Vision – ECCV 2024, Springer Nature Switzer- land, Cham

    Open- vocabulary sam: Segment and recognize twenty-thousand classes interac- tively, in: Leonardis, A., Ricci, E., Roth, S., Russakovsky, O., Sattler, T., Varol, G. (Eds.), Computer Vision – ECCV 2024, Springer Nature Switzer- land, Cham. pp. 419–437. Zareian, A., Rosa, K.D., ...

    2024

  5. [25]

    14388–14397

    Open-vocabulary ob- ject detection using captions, in: 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pp. 14388–14397. Zhang, F., Zhou, T., Li, B., He, H., Ma, C., Zhang, T., Yao, J., Zhang, Y ., Wang, Y .,

    2021

  6. [26]

    (Eds.), Advances in Neural Information Process- ing Systems, Curran Associates, Inc

    Uncovering prototypical knowledge for weakly open-vocabulary semantic segmentation, in: Oh, A., Naumann, T., Globerson, A., Saenko, K., Hardt, M., Levine, S. (Eds.), Advances in Neural Information Process- ing Systems, Curran Associates, Inc.. pp. 73652–73665. URL:https: //pro...

    2023

  7. [27]

    2593–2602

    Zero-shot instance seg- mentation, in: 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pp. 2593–2602. Zhou, T., Xia, W., Zhang, F., Chang, B., Wang, W., Yuan, Y ., Konukoglu, E., Cremers, D.,

    2021

  8. [28]

    URL:https://arxiv.org/abs/2408.12957,arXiv:2408.12957

    Image segmentation in foundation model era: A survey. URL:https://arxiv.org/abs/2408.12957,arXiv:2408.12957. Zhu, Z., Wei, Y ., Wang, J., Gan, Z., Zhang, Z., Wang, L., Hua, G., Wang, L., Liu, Z., Hu, H.,

    arXiv

  9. [29]

    arXiv preprint arXiv:2211.11694

    Exploring discrete diffusion models for image captioning. arXiv preprint arXiv:2211.11694 . 13 Zou, X., Dou, Z.Y ., Yang, J., Gan, Z., Li, L., Li, C., Dai, X., Behl, H., Wang, J., Yuan, L., Peng, N., Wang, L., Lee, Y .J., Gao, J.,

    arXiv

  10. [30]

    15116–15127

    Generalized decoding for pixel, image, and language, in: 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pp. 15116–15127. doi:10.1109/CVPR52729.2023.01451. 14

    2023

  11. [2014]

    Microsoft coco: Common objects in context, in: Com- puter Vision–ECCV 2014: 13th European Conference, Zurich, Switzerland, September 6-12, 2014, Proceedings, Part V 13, Springer. pp. 740–755. Liu, H., Li, C., Wu, Q., Lee, Y .J.,

    2014

  12. [2019]

    BERT: Pre-training of deep bidirectional transformers for language understanding, in: Burstein, J., Doran, C., Solorio, T. (Eds.), Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, V ...

    2019

  13. [2020]

    International journal of computer vision 128, 1956–1981

    The open images dataset v4: Unified image classification, object detection, and visual relationship detection at scale. International journal of computer vision 128, 1956–1981. Li, H., Su, J., Zhou, D., Cao, M.,

    1956

  14. [2021]

    (Eds.), Advances in Neural Informa- tion Processing Systems, Curran Associates, Inc

    Per-pixel classification is not all you need for semantic segmentation, in: Ranzato, M., Beygelzimer, A., Dauphin, Y ., Liang, P., Vaughan, J.W. (Eds.), Advances in Neural Informa- tion Processing Systems, Curran Associates, Inc.. pp. 17864–17875. URL: https://proceedings.neur...

    2021

  15. [2022]

    1280–1289

    Masked-attention mask transformer for universal image segmentation, in: 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pp. 1280–1289. doi:10.1109/CVPR52688.2022.00135. Cheng, B., Schwing, A., Kirillov, A.,

    2022

  16. [2023]

    3992–4003

    Segment anything, in: 2023 IEEE/CVF International Conference on Computer Vision (ICCV), pp. 3992–4003. doi:10.1109/ICCV51070.2023.00371. Kuznetsova, A., Rom, H., Alldrin, N., Uijlings, J., Krasin, I., Pont-Tuset, J., Kamali, S., Popov, S., Malloci, M., Kolesnikov, A., et al.,

    2023

  17. [2024]

    27948–27959

    Omg-seg: Is one model good enough for all segmentation?, in: 2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pp. 27948–27959. doi:10.1109/CVPR52733.2024.02640. Liang, J.C., Zhou, T., Liu, D., Wang, W.,

    2024

  18. [2025]

    Neurocomput- ing 630, 129702

    Image–text aggregation for open-vocabulary semantic segmentation. Neurocomput- ing 630, 129702. URL:https://www.sciencedirect.com/science/ article/pii/S0925231225003741, doi:https://doi.org/10.1016/ j.neucom.2025.129702. Chiang, W.L., Li, Z., Lin, Z., Sheng, Y ., Wu, Z., Zhang...

    2025

  19. [2026]

    Neurocomputing 660, 131844

    Fa-seg: A fast and accurate diffusion- based method for open-vocabulary segmentation. Neurocomputing 660, 131844. URL:https://www.sciencedirect.com/science/ article/pii/S0925231225025160, doi:https://doi.org/10.1016/ j.neucom.2025.131844. Chen, L., Yang, Q., Ding, K., Li, Z., ...

Pith tools

Reviewed August 16, 2026 · model on record in the stance chip above.