Pith. sign in

REVIEW 3 major objections 4 minor 72 references

Through the Looking Glass: A Dual Perspective on Weakly-Supervised Few-Shot Segmentation

T0 review · 3 major / 4 minor · reviewed 2026-08-05 · 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 treating support and query images as complementary, asymmetric views of the same category—rather than feeding them through identical network branches—lets a weakly supervised few-shot segmentation model with about 1/24

desk verdict Novel heterogeneous-branch idea with a strong but imperfectly supported empirical claim; the parameter estimation needs fixing and the COCO number conflicts. read the letter →

arxiv 2508.16159 v2 pith:P32W2YPL submitted 2025-08-22 cs.CV cs.AI

Jiaqi Ma , Guo-Sen Xie , Fang Zhao , Zechao Li This is my paper · ORCID
classification cs.CVcs.AI
keywords weakly-supervisedfew-shotsemanticsegmentationimage-levellabelsheterogeneousmeta-learningCLIPoptimaltransportsupport-querydualbrancheslightweight
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 claims that the standard meta-learning recipe for few-shot semantic segmentation—running the support image and the query image through identical network branches—discards useful signal by over-homogenizing the two views. It proposes TLG, a 'homologous but heterogeneous' network in which support and query are processed by asymmetric branches at different feature depths, then recombined by three modules: heterogeneous visual aggregation (HA), heterogeneous transfer (HT), and heterogeneous CLIP (HC). The reported results are large: 13.2% improvement on Pascal-5i and 9.7% on COCO-20i over prior weakly supervised state of the art, with around 1/24 of the parameters (4.47M versus 108.35M for AFANet). The authors further claim TLG is the first image-level-only (weakly supervised) segmentation model to beat fully supervised models that use the same backbone. If true, the paper's core message is that support-query pairs should be treated as complementary perspectives rather than interchangeable examples.

What carries the argument

The load-bearing object is the 'homologous but heterogeneous' dual-branch architecture: the support branch and query branch are not weight-shared at the same layer; each branch is assigned a different fixed range of backbone feature layers, so the network sees the same category through two complementary lenses. The HA module preserves semantic commonality while exchanging complementary features; the HT module uses an optimal-transport-based context-attention mechanism to remove semantic noise; the HC module fuses CLIP text embeddings as a class-level anchor. The asymmetry itself is the mechanism: it forces the two perspectives to carry different information that the later fusion can exploit.

What would settle it

Run TLG with a single shared backbone plus only the Heterogeneous CLIP module, deleting HA and HT: if accuracy stays near the full model, heterogeneity is not the cause of the reported gain. Alternatively, shuffle or randomize the fixed support-query layer assignments across many runs on Pascal-5i: if mean IoU barely changes, the specific layer asymmetry is not load-bearing.

Watch

Extended reading notes

Core claim

On its own terms, the paper's central claim is that the semantic homology of a support-query pair—same coarse category, related attributes—does not justify shared network weights. TLG instead assigns the two branches fixed but different backbone layers, so that one captures low-level, texture-dominant features and the other captures high-level semantic features. The Heterogeneous Visual Aggregation module aligns and exchanges these complementary features; the Heterogeneous Transfer module applies an optimal-transport context-attention mechanism to suppress noise and sharpen boundaries; the Heterogeneous CLIP module injects textual embeddings from CLIP as a stable class anchor. The paper repo

Load-bearing premise

The paper's key premise is that its gains come from deliberately heterogeneous support/query branches; that premise depends on a fixed, hand-chosen layer assignment that never adapts to the input, and if the real driver is the CLIP prior or optimistic parameter estimates in the comparison, the dual-perspective explanation loses its force.

Editorial extensions

If this is right

  • If the reported results hold, image-level class labels can substitute for pixel-level masks in few-shot segmentation, removing the most expensive annotation step.
  • At 4.47M parameters and with one-shot inference FLOPs of 41.25G on ResNet-50, the model is small and fast enough to make weakly supervised few-shot segmentation a plausible edge-deployment option.
  • The 'two views of one category' principle suggests other meta-learning tasks—object detection, recognition, retrieval—could benefit from asymmetric support/query branches instead of tied weights.
  • Winning against fully supervised VRP-SAM and PGMA-Net under the same backbone resets the baseline conversation: weak supervision can no longer be assumed to be intrinsically inferior.
  • The HC module shows that a frozen CLIP text encoder can be injected into a few-shot segmentation pipeline at negligible parameter cost, pointing toward prompt-based extensions.

Reading between the lines

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

  • If the paper's gain were decomposed, the CLIP prior may account for a substantial share; the authors' ablations need to separate 'heterogeneity' from 'extra text guidance' before the dual-perspective mechanism can be credited with the full margin.
  • The fixed support-query layer assignments are manually chosen and input-independent (a limitation the authors state); a learnable, data-dependent layer selection is a natural next step and could either push the numbers higher or expose where the current assignment fails.
  • Parameter-count claims for several comparison methods (Siam et al., Zhang et al., MIAPNet) come from replicated experiments rather than official releases, so the exact '1/24' factor is approximate and should be read with that caveat.
  • The abstract reports a 9.7% gain on COCO-20i while the full text at one point states 7.9%; a reader should confirm which number corresponds to the final evaluation protocol before treating the margin as settled.
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 proposes TLG, a weakly-supervised few-shot semantic segmentation method built on a 'homologous but heterogeneous' dual-branch design: the support and query branches use different network configurations (HA modules), a heterogeneous transport (HT) module suppresses noise, and heterogeneous CLIP (HC) textual information is injected for multimodal generalization. The authors claim, with 1/24 of the parameters of existing state-of-the-art models, a 13.2% improvement on Pascal-5i and (depending on the version of the abstract) 7.9% or 9.7% on COCO-20i, and further claim to be the first weakly supervised image-level model to outperform fully supervised pixel-level models under the same backbone. The manuscript includes qualitative visualizations, a parameter-efficiency analysis, some FLOPs/latency numbers, and a limitations section that acknowledges the fixed, manually designed support-query layer assignments. The code is openly available.

Significance. If the reported numbers are correct, the result is significant: it would demonstrate a parameter-efficient weakly supervised few-shot segmentation model that surpasses fully supervised counterparts under the same backbone, supporting the paper's central thesis that heterogeneous support/query branches improve semantic complementarity. The paper also provides a public code release and reports inference-cost metrics, which are useful for the community. However, the significance currently rests on load-bearing empirical claims that are not fully auditable: the parameter-count comparison uses an ad hoc estimator for several baselines, the COCO-20i improvement is quoted inconsistently, and the key experimental tables are absent from the version under review. These issues must be addressed before the claims can be accepted.

major comments (3)
  1. [§V (Lightweighting Analysis)] The parameter counts for Siam et al., Zhang et al., and MIAPNet are computed as 'the average parameter count from our replication experiments and TLG.' This is not a valid estimator of a baseline's parameter count: averaging the unknown baseline count with TLG's 4.47M pulls the estimate toward TLG and can only flatter the reported ratio. For example, if a baseline actually had 20M parameters, the midpoint with TLG would be ~12.2M, giving a parameter ratio of ~2.7×, not 24×. Since the abstract's '1/24 of the parameters' headline depends on this calculation, the central lightweighting claim needs to be re-supported with actual parameter counts from official implementations or exact replications of each baseline.
  2. [Abstract and experimental tables] The abstract in the full text reports a 7.9% improvement on COCO-20i, while the version of the abstract quoted in the arXiv metadata states 9.7%. Moreover, Tables I, II, IV, and V, which are the only sources for the Pascal-5i/COCO-20i comparisons, the fully-supervised comparison, the ablation results, and the FLOPs/latency analysis, are not present in the manuscript under review. The prose in §J reports only a few isolated numbers (e.g., TLG 77.2 vs. VRP-SAM 71.9 and PGMA-Net 74.1). Without the complete tables and a single consistent set of headline figures, the central performance claims cannot be audited. This is a major issue for an empirical paper.
  3. [§VI and attribution of gains] The authors acknowledge in §VI that the heterogeneous layer selection is manually designed and fixed across inputs. Because the paper's explanatory claim is that heterogeneity itself (rather than, say, the CLIP prior or the HT module) drives the improvements, the absence of any sensitivity analysis or ablation over alternative layer assignments is a load-bearing gap. The qualitative ablation in Fig. 9 is suggestive but not quantitative. The authors should report ablation numbers that vary the fixed layer assignments (e.g., support/query swapping, different layer depths) and isolate the contribution of the HA/HT/HC modules to the reported Pascal-5i and COCO-20i improvements.
minor comments (4)
  1. [Abstract and §I] The module name is inconsistent: the abstract says 'heterogeneous transfer (HT)' while the introduction and Fig. 9 say 'Heterogeneous Transport (HT).' Please use one name throughout.
  2. [§I] There are minor formatting typos, such as 'Through theLookingGlass' (missing space) and the sentence ending 'namedTLG.' — please fix.
  3. [§V] The phrase 'with only 1/24 of the parameters of existing state-of-the-art models' is ambiguous because the comparison in §V is only against AFANet. Please specify which baseline(s) the 1/24 figure refers to.
  4. [Fig. 9] The qualitative caption states 'mean IoU = 0.9' for one visualization; it is unclear what this number represents (e.g., the IoU is shown in yellow top-right, not a mean over a dataset). Please clarify.

Circularity Check

1 steps flagged · score 2.0 of 10

No definitional circularity in the core method; one self-referential parameter-count estimate weakens the lightweighting comparison but does not drive the main accuracy claims.

  1. other [Section V (Lightweighting Analysis), parameter-count paragraph]
    "The learnable parameters for IMR-HSNet and AFANet are derived from experimental replication, while for Siam et al., Zhang et al., and MIAPNet, whose models are not open-source, we use the average parameter count from our replication experiments and TLG."

    The baseline parameter count for three non-open-source methods is defined as the average of the authors' own replication and TLG's own parameter count. Consequently, any parameter-reduction ratio involving these baselines is partly constructed from TLG itself rather than being an independent measurement. Since TLG's 4.47M count is far smaller than typical published FSS models, averaging it into the baseline estimates inflates the apparent lightweighting gain by construction for these three methods. This does not affect the segmentation-accuracy comparisons, which are standard empirical evaluations, but it makes the parameter-efficiency comparison self-referential for those particular baselines.

full rationale

TLG is an empirical architecture paper, not a mathematical derivation. The central claims are measured accuracies on Pascal-5i and COCO-20i against published baselines, and there is no fitted parameter that is later relabeled as a prediction, no uniqueness theorem imported from the authors, and no ansatz smuggled in via citation. The use of AFANet, the authors' own prior model, as a state-of-the-art baseline is a self-citation but not load-bearing: AFANet is a published, externally accessible method and the comparison is empirical. The only circularity-adjacent issue is the Lightweighting Analysis, where three non-open-source baselines are assigned parameter counts equal to the average of the authors' replication and TLG's own count. That procedure makes the '1/24 parameters' claim partly self-referential for those three baselines, because the baseline count is partially defined by the model it is compared against. However, the headline '1/24' is presented specifically against AFANet's 108.35M count, which comes from replication rather than from averaging in TLG. The paper also acknowledges the fixed, manually chosen layer selection as a limitation, which is a legitimate weakness but not circularity. Missing tables in the provided text prevent full auditing of every comparison, but that is an omission of evidence rather than a circular reduction. Overall, the core contribution is a heterogeneous dual-branch network evaluated empirically, so no significant definitional circularity is present; the self-referential parameter-count estimate warrants a small score increase.

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

The central claim rests on a manually designed architecture with fixed layer assignments and ad hoc prompts. The three core modules (HA, HT, HC) are the invented entities, none of which have independent evidence outside this paper. The paper contributes an architecture design rather than a parameter-free derivation, and the visible text does not disclose all tuning choices.

free parameters (3)
  • Fixed support-query layer assignments = unknown (manually chosen)
    The heterogeneous layer selection strategy is manually designed with fixed assignments that remain unchanged across inputs, as stated in Section VI (Limitations).
  • Foreground/background prompts for CLIP = fixed prompts (not specified)
    Section VI mentions fixed foreground-background prompts; these are chosen ad hoc and used by the HC module.
  • Configuration of HA/HT/HC modules = not specified in visible text
    The architecture hyperparameters such as layer counts, attention heads, and module dimensions are not visible in the provided excerpt, and given the manually designed structure, they are tuned choices.
assumptions (3)
  • domain assumption Support and query images of the same category share semantic homology
    Stated in the intro: support/query pairs are homologous (same coarse category) but heterogeneous (fine attributes). This underpins the whole design.
  • ad hoc to paper Identical network architectures cause over-semantic homogenization that limits performance
    This is the paper's core motivation, asserted in the intro and not proven. It justifies the heterogeneous design and is load-bearing for the claimed improvement.
  • ad hoc to paper The fixed manual layer assignment is sufficient for good performance
    The Limitations section itself says this is a limitation: the strategy is manually designed and fixed, limiting the model's adaptability. The paper's success depends on this fixed choice being adequate.
invented entities (3)
  • Heterogeneous visual aggregation (HA) modules
    purpose: Enhance complementarity while preserving semantic commonality between support and query branches
    Only evaluated within this paper; no external falsifiable handle or benchmark outside this work.
  • Heterogeneous transport (HT) module
    purpose: Reduce semantic noise via context-attention based on optimal transport
    An in-paper module; no external validation provided.
  • Heterogeneous CLIP (HC) textual information
    purpose: Enhance generalization using multimodal CLIP knowledge
    Built on existing CLIP, but the specific heterogeneous text fusion is only tested here.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Through the Looking Glass: A Dual Perspective on Weakly-Supervised Few-Shot Segmentation." pith.science (2026). https://pith.science/paper/P32W2YPL

@misc{pith2026250816159,
  author       = {Pith},
  title        = {Pith review of: Through the Looking Glass: A Dual Perspective on Weakly-Supervised Few-Shot Segmentation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/P32W2YPL}},
  note         = {Machine review of arXiv:2508.16159}
}
read the original abstract

Meta-learning aims to uniformly sample homogeneous support-query pairs, characterized by the same categories and similar attributes, and extract useful inductive biases through identical network architectures. However, this identical network design results in over-semantic homogenization. To address this, we propose a novel homologous but heterogeneous network. By treating support-query pairs as dual perspectives, we introduce heterogeneous visual aggregation (HA) modules to enhance complementarity while preserving semantic commonality. To further reduce semantic noise and amplify the uniqueness of heterogeneous semantics, we design a heterogeneous transfer (HT) module. Finally, we propose heterogeneous CLIP (HC) textual information to enhance the generalization capability of multimodal models. In the weakly-supervised few-shot semantic segmentation (WFSS) task, with only 1/24 of the parameters of existing state-of-the-art models, TLG achieves a 13.2\% improvement on Pascal-5\textsuperscript{i} and a 9.7\% improvement on COCO-20\textsuperscript{i}. To the best of our knowledge, TLG is also the first weakly supervised (image-level) model that outperforms fully supervised (pixel-level) models under the same backbone architectures. The code is available at https://github.com/jarch-ma/TLG.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

72 extracted references · 60 canonical work pages

  1. [1]

    Training language models to follow instructions with human feedback,

    L. Ouyang, J. Wu, X. Jiang, D. Almeida, C. Wainwright, P. Mishkin, C. Zhang, S. Agarwal, K. Slama, A. Rayet al., “Training language models to follow instructions with human feedback,”Advances in neural information processing systems, vol. 35, pp. 27 730–27 744, 2022

    2022

  2. [2]

    Segment anything,

    A. Kirillov, E. Mintun, N. Ravi, H. Mao, C. Rolland, L. Gustafson, T. Xiao, S. Whitehead, A. C. Berg, W.-Y . Loet al., “Segment anything,” inProceedings of the IEEE/CVF International Conference on Computer Vision, 2023, pp. 4015–4026

    2023

  3. [3]

    Learning transferable visual models from natural language supervision,

    A. Radford, J. W. Kim, C. Hallacy, A. Ramesh, G. Goh, S. Agarwal, G. Sastry, A. Askell, P. Mishkin, J. Clarket al., “Learning transferable visual models from natural language supervision,” inInternational conference on machine learning, 2021, pp. 8748–8763

    2021

  4. [4]

    Llama: Open and efficient foundation language models,

    H. Touvron, T. Lavril, G. Izacard, X. Martinet, M.-A. Lachaux, T. Lacroix, B. Rozi `ere, N. Goyal, E. Hambro, F. Azharet al., “Llama: Open and efficient foundation language models,”arXiv preprint arXiv:2302.13971, 2023

    arXiv 2023

  5. [5]

    NP-SemiSeg: When neural processes meet semi-supervised semantic segmentation,

    J. Wang, D. Massiceti, X. Hu, V . Pavlovic, and T. Lukasiewicz, “NP-SemiSeg: When neural processes meet semi-supervised semantic segmentation,” inProceedings of the 40th International Conference on Machine Learning, vol. 202, 23–29 Jul 2023, pp. 36 138–36 156

    work page 2023

  6. [6]

    Meta-learning in neural networks: A survey,

    T. Hospedales, A. Antoniou, P. Micaelli, and A. Storkey, “Meta-learning in neural networks: A survey,”IEEE transactions on pattern analysis and machine intelligence, vol. 44, no. 9, pp. 5149–5169, 2021

    2021

  7. [7]

    Knowledge-guided semantic transfer network for few-shot image recognition,

    Z. Li, H. Tang, Z. Peng, G.-J. Qi, and J. Tang, “Knowledge-guided semantic transfer network for few-shot image recognition,”IEEE Trans- actions on Neural Networks and Learning Systems, 2023

    work page 2023

  8. [8]

    Few-shot image recognition with knowledge transfer,

    Z. Peng, Z. Li, J. Zhang, Y . Li, G.-J. Qi, and J. Tang, “Few-shot image recognition with knowledge transfer,” inProceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), October 2019

    work page 2019

Show all 72 references
  1. [9]

    Few-shot cross-domain object detection with instance-level prototype-based meta-learning,

    L. Zhang, B. Zhang, B. Shi, J. Fan, and T. Chen, “Few-shot cross-domain object detection with instance-level prototype-based meta-learning,” IEEE Transactions on Circuits and Systems for Video Technology, vol. 34, no. 10, pp. 9078–9089, 2024

    2024

  2. [10]

    Mul-sno: a novel prediction tool for s-nitrosylation sites based on deep learning methods,

    Q. Zhao, J. Ma, Y . Wang, F. Xie, Z. Lv, Y . Xu, H. Shi, and K. Han, “Mul-sno: a novel prediction tool for s-nitrosylation sites based on deep learning methods,”IEEE Journal of Biomedical and Health Informatics, vol. 26, no. 5, pp. 2379–2387, 2021

    2021

  3. [11]

    Deep collaborative embedding for social image understanding,

    Z. Li, J. Tang, and T. Mei, “Deep collaborative embedding for social image understanding,”IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 41, no. 9, pp. 2070–2083, 2019

  4. [12]

    Medical image segmentation review: The success of u-net,

    R. Azad, E. K. Aghdam, A. Rauland, Y . Jia, A. H. Avval, A. Bozorgpour, S. Karimijafarbigloo, J. P. Cohen, E. Adeli, and D. Merhof, “Medical image segmentation review: The success of u-net,”IEEE Transactions on Pattern Analysis and Machine Intelligence, 2024. JOURNAL OF LATEX ...

    2024

  5. [13]

    Slowfast diversity-aware prototype learning for egocentric action recognition,

    G. Dai, X. Shu, R. Yan, P. Huang, and J. Tang, “Slowfast diversity-aware prototype learning for egocentric action recognition,” inProceedings of the ACM International Conference on Multimedia (ACM MM), 2023, pp. 7549–7558

    2023

  6. [14]

    Gpt4ego: Unleashing the potential of pre-trained models for zero-shot egocentric action recognition,

    G. Dai, X. Shu, W. Wu, R. Yan, and J. Zhang, “Gpt4ego: Unleashing the potential of pre-trained models for zero-shot egocentric action recognition,”IEEE Transactions on Multimedia, vol. 27, pp. 401–413, 2025

    2025

  7. [15]

    Thinkmat- ter: Panoramic-aware instructional semantics for monocular vision-and- language navigation,

    G. Dai, S. Wang, H. Zhao, B. Zhu, Q. Sun, and X. Shu, “Thinkmat- ter: Panoramic-aware instructional semantics for monocular vision-and- language navigation,”IEEE Transactions on Image Processing (TIP), vol. 35, pp. 1937–1950, 2026

    1937

  8. [16]

    Survey of single image super-resolution reconstruction,

    K. Li, S. Yang, R. Dong, X. Wang, and J. Huang, “Survey of single image super-resolution reconstruction,”IET Image Processing, vol. 14, no. 11, pp. 2273–2290, 2020

    2020

  9. [17]

    An efficient encoder-decoder architecture with top-down attention for speech separation,

    K. Li, R. Yang, and X. Hu, “An efficient encoder-decoder architecture with top-down attention for speech separation,” inICLR, 2023

    2023

  10. [18]

    The genomic basis of adaptive evolution in threespine sticklebacks,

    F. C. Jones, M. G. Grabherr, Y . F. Chan, P. Russell, E. Mauceli, J. Johnson, R. Swofford, M. Pirun, M. C. Zody, S. Whiteet al., “The genomic basis of adaptive evolution in threespine sticklebacks,”Nature, vol. 484, no. 7392, pp. 55–61, 2012

    2012

  11. [19]

    Context disentangling and prototype inheriting for robust visual grounding,

    W. Tang, L. Li, X. Liu, L. Jin, J. Tang, and Z. Li, “Context disentangling and prototype inheriting for robust visual grounding,”IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 46, no. 5, pp. 3213– 3229, 2023

    2023

  12. [20]

    Ssp-sam: Sam with semantic-spatial prompt for referring expression segmentation,

    W. Tang, X. Liu, Y . Sun, and Z. Li, “Ssp-sam: Sam with semantic-spatial prompt for referring expression segmentation,”IEEE Transactions on Circuits and Systems for Video Technology, vol. 36, no. 5, pp. 6374– 6389, 2026

    2026

  13. [21]

    Efficient audio-visual speech separation with discrete lip semantics and multi-scale global-local attention,

    K. Li, K. Gao, and X. Hu, “Efficient audio-visual speech separation with discrete lip semantics and multi-scale global-local attention,” in International Conference on Learning Representations (ICLR), 2026

    2026

  14. [22]

    Carroll,Through the Looking-Glass: and What Alice Saw There

    L. Carroll,Through the Looking-Glass: and What Alice Saw There. Digital Scanning Inc, 2007

    2007

  15. [23]

    Hyperbolic insights with knowledge distillation for cross-domain few-shot learning,

    X. Yang, D. Kong, N. Wang, and X. Gao, “Hyperbolic insights with knowledge distillation for cross-domain few-shot learning,”IEEE Trans- actions on Image Processing, vol. 34, pp. 1921–1933, 2025

    1921

  16. [24]

    A two-stage approach to few-shot learning for image recognition,

    D. Das and C. S. G. Lee, “A two-stage approach to few-shot learning for image recognition,”IEEE Transactions on Image Processing, vol. 29, pp. 3336–3350, 2020

    2020

  17. [25]

    Enhanc- ing information maximization with distance-aware contrastive learning for source-free cross-domain few-shot learning,

    H. Xu, L. Liu, S. Zhi, S. Fu, Z. Su, M.-M. Cheng, and Y . Liu, “Enhanc- ing information maximization with distance-aware contrastive learning for source-free cross-domain few-shot learning,”IEEE Transactions on Image Processing, vol. 33, pp. 2058–2073, 2024

  18. [26]

    Generalized odin: Detecting out-of-distribution image without learning from out-of-distribution data,

    Y .-C. Hsu, Y . Shen, H. Jin, and Z. Kira, “Generalized odin: Detecting out-of-distribution image without learning from out-of-distribution data,” inProceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2020, pp. 10 951–10 960

    2020

  19. [27]

    Sequence homology score-based deep fuzzy network for identifying therapeutic peptides,

    X. Guo, Z. Zheng, K. H. Cheong, Q. Zou, P. Tiwari, and Y . Ding, “Sequence homology score-based deep fuzzy network for identifying therapeutic peptides,”Neural Networks, vol. 178, p. 106458, 2024

    2024

  20. [28]

    Seg4diff: Unveiling open-vocabulary semantic segmentation in text-to-image diffusion transformers,

    C. Kim, H. Shin, E. Hong, H. Yoon, A. Arnab, P. H. Seo, S. Hong, and S. Kim, “Seg4diff: Unveiling open-vocabulary semantic segmentation in text-to-image diffusion transformers,”Advances in Neural Information Processing Systems, vol. 38, pp. 71 685–71 724, 2026

    2026

  21. [29]

    Exploring efficient open-vocabulary segmentation in the remote sensing,

    B. Li, H. Dong, D. Zhang, Z. Zhao, H. Sun, and J. Gao, “Exploring efficient open-vocabulary segmentation in the remote sensing,” inPro- ceedings of the AAAI Conference on Artificial Intelligence, vol. 40, no. 8, 2026, pp. 5982–5991

    2026

  22. [30]

    Learning attention-guided pyrami- dal features for few-shot fine-grained recognition,

    H. Tang, C. Yuan, Z. Li, and J. Tang, “Learning attention-guided pyrami- dal features for few-shot fine-grained recognition,”Pattern Recognition, vol. 130, p. 108792, 2022

    2022

  23. [31]

    Recent advances in predicting protein s- nitrosylation sites,

    Q. Zhao, J. Ma, F. Xie, Y . Wang, Y . Zhang, H. Li, Y . Sun, L. Wang, M. Guo, and K. Han, “Recent advances in predicting protein s- nitrosylation sites,”BioMed Research International, vol. 2021, no. 1, p. 5542224, 2021

    2021

  24. [32]

    Cross-modal contrastive learning network for few-shot action recognition,

    X. Wang, Y . Yan, H.-M. Hu, B. Li, and H. Wang, “Cross-modal contrastive learning network for few-shot action recognition,”IEEE Transactions on Image Processing, vol. 33, pp. 1257–1271, 2024

    2024

  25. [33]

    Opmapper: Enhancing open-vocabulary semantic segmentation with multi-guidance information,

    X. Wang, C. Si, X. Yang, Y . Zhao, W. Wang, X. Yang, and W. Shen, “Opmapper: Enhancing open-vocabulary semantic segmentation with multi-guidance information,”Advances in Neural Information Process- ing Systems, vol. 38, pp. 143 331–143 358, 2026

    2026

  26. [34]

    A closer look at the few-shot adaptation of large vision-language models,

    J. Silva-Rodr ´ıguez, S. Hajimiri, I. Ben Ayed, and J. Dolz, “A closer look at the few-shot adaptation of large vision-language models,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), June 2024, pp. 23 681–23 690

    2024

  27. [35]

    Advances and challenges in meta-learning: A technical re- view,

    A. Vettoruzzo, M.-R. Bouguelia, J. Vanschoren, T. R ¨ognvaldsson, and K. Santosh, “Advances and challenges in meta-learning: A technical re- view,”IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 46, no. 7, pp. 4763–4779, 2024

    2024

  28. [36]

    Zero-shot neural architecture search: Challenges, solutions, and opportunities,

    G. Li, D. Hoang, K. Bhardwaj, M. Lin, Z. Wang, and R. Mar- culescu, “Zero-shot neural architecture search: Challenges, solutions, and opportunities,”IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 46, no. 12, pp. 7618–7635, 2024

    2024

  29. [37]

    Tapb: an interventional debiasing framework for alleviating target prior bias in drug-target interaction prediction,

    G. Lin, X. Zhang, Z. Ren, Q. Zou, P. Tiwari, C. Zhou, and Y . Ding, “Tapb: an interventional debiasing framework for alleviating target prior bias in drug-target interaction prediction,”Nature Communications, 2025

    2025

  30. [38]

    Heteroge- neous graph neural network,

    C. Zhang, D. Song, C. Huang, A. Swami, and N. V . Chawla, “Heteroge- neous graph neural network,” inProceedings of the 25th ACM SIGKDD international conference on knowledge discovery & data mining, 2019, pp. 793–803

    2019

  31. [39]

    Cross-heterogeneity graph few-shot learning,

    P. Ding, Y . Wang, and G. Liu, “Cross-heterogeneity graph few-shot learning,” inProceedings of the 32nd ACM International Conference on Information and Knowledge Management, 2023, pp. 420–429

    2023

  32. [40]

    Weakly supervised object detection using proposal- and semantic-level relationships,

    D. Zhang, W. Zeng, J. Yao, and J. Han, “Weakly supervised object detection using proposal- and semantic-level relationships,”IEEE Trans- actions on Pattern Analysis and Machine Intelligence, vol. 44, no. 6, pp. 3349–3363, 2022

    2022

  33. [41]

    Improving pedestrian attribute recognition with weakly-supervised multi-scale attribute-specific local- ization,

    C. Tang, L. Sheng, Z. Zhang, and X. Hu, “Improving pedestrian attribute recognition with weakly-supervised multi-scale attribute-specific local- ization,” inProceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), October 2019

    2019

  34. [42]

    Weakly supervised object localization and detection: A survey,

    D. Zhang, J. Han, G. Cheng, and M.-H. Yang, “Weakly supervised object localization and detection: A survey,”IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 44, no. 9, pp. 5866–5885, 2022

    2022

  35. [43]

    Scale-aware graph neural network for few-shot semantic segmentation,

    G.-S. Xie, J. Liu, H. Xiong, and L. Shao, “Scale-aware graph neural network for few-shot semantic segmentation,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), June 2021, pp. 5475–5484

    2021

  36. [44]

    Few-shot se- mantic segmentation with cyclic memory network,

    G.-S. Xie, H. Xiong, J. Liu, Y . Yao, and L. Shao, “Few-shot se- mantic segmentation with cyclic memory network,” inProceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), October 2021, pp. 7293–7302

    2021

  37. [45]

    Semantically meaningful class prototype learning for one-shot image segmentation,

    T. Chen, G.-S. Xie, Y . Yao, Q. Wang, F. Shen, Z. Tang, and J. Zhang, “Semantically meaningful class prototype learning for one-shot image segmentation,”IEEE Transactions on Multimedia, vol. 24, pp. 968–980, 2022

    2022

  38. [46]

    A self-supervised-driven open-set un- supervised domain adaptation method for optical remote sensing image scene classification and retrieval,

    S. Wang, D. Hou, and H. Xing, “A self-supervised-driven open-set un- supervised domain adaptation method for optical remote sensing image scene classification and retrieval,”IEEE Transactions on Geoscience and Remote Sensing, vol. 61, pp. 1–15, 2023

    2023

  39. [47]

    Weakly supervised one shot segmentation,

    H. Raza, M. Ravanbakhsh, T. Klein, and M. Nabi, “Weakly supervised one shot segmentation,” inProceedings of the IEEE/CVF International Conference on Computer Vision (ICCV) Workshops, Oct 2019

    2019

  40. [48]

    Weakly supervised few-shot object segmentation using co- attention with visual and semantic embeddings,

    M. Siam, N. Doraiswamy, B. N. Oreshkin, H. Yao, and M. Jager- sand, “Weakly supervised few-shot object segmentation using co- attention with visual and semantic embeddings,”arXiv preprint arXiv:2001.09540, 2020

    2001 arXiv

  41. [49]

    A pixel-level meta-learner for weakly supervised few-shot semantic segmentation,

    Y .-H. Lee, F.-E. Yang, and Y .-C. F. Wang, “A pixel-level meta-learner for weakly supervised few-shot semantic segmentation,” inProceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, 2022, pp. 2170–2180

    2022

  42. [50]

    Weakly-supervised object represen- tation learning for few-shot semantic segmentation,

    X. Ying, X. Li, and M. C. Chuah, “Weakly-supervised object represen- tation learning for few-shot semantic segmentation,” inProceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, 2021, pp. 1497–1506

    2021

  43. [51]

    Iterative few-shot semantic segmentation from image label text,

    H. Wang, L. Liu, W. Zhang, J. Zhang, Z. Gan, Y . Wang, C. Wang, and H. Wang, “Iterative few-shot semantic segmentation from image label text,”arXiv preprint arXiv:2303.05646, 2023

    2023 arXiv

  44. [52]

    Weakly supervised few-shot and zero-shot semantic segmentation with mean instance aware prompt learning,

    P. Pandey, M. Chasmai, M. Natarajan, and B. Lall, “Weakly supervised few-shot and zero-shot semantic segmentation with mean instance aware prompt learning,” in2023 IEEE International Conference on Multimedia and Expo (ICME). IEEE, 2023, pp. 1–6

    2023

  45. [53]

    Afanet: Adaptive frequency-aware network for weakly-supervised few-shot semantic segmentation,

    J. Ma, G.-S. Xie, F. Zhao, and Z. Li, “Afanet: Adaptive frequency-aware network for weakly-supervised few-shot semantic segmentation,”IEEE Transactions on Multimedia, vol. 27, pp. 4018–4028, 2025

    2025

  46. [54]

    Learning deep features for discriminative localization,

    B. Zhou, A. Khosla, A. Lapedriza, A. Oliva, and A. Torralba, “Learning deep features for discriminative localization,” inProceedings of the IEEE conference on computer vision and pattern recognition, 2016, pp. 2921– 2929. JOURNAL OF LATEX CLASS FILES, VOL. 14, NO. 8, AUGUST 2021 14

    2016

  47. [55]

    Hsnet: A hybrid semantic network for polyp segmentation,

    W. Zhang, C. Fu, Y . Zheng, F. Zhang, Y . Zhao, and C.-W. Sham, “Hsnet: A hybrid semantic network for polyp segmentation,”Computers in biology and medicine, vol. 150, p. 106173, 2022

    2022

  48. [56]

    Attention is all you need,

    A. Vaswani, N. Shazeer, N. Parmar, J. Uszkoreit, L. Jones, A. N. Gomez, L. u. Kaiser, and I. Polosukhin, “Attention is all you need,” vol. 30. Curran Associates, Inc., 2017

    2017

  49. [57]

    Generalized few-shot 3d point cloud segmentation with vision-language model,

    Z. An, G. Sun, Y . Liu, R. Li, J. Han, E. Konukoglu, and S. Belongie, “Generalized few-shot 3d point cloud segmentation with vision-language model,” inProceedings of the Computer Vision and Pattern Recognition Conference, 2025, pp. 16 997–17 007

    2025

  50. [58]

    Deep visual attention prediction,

    W. Wang and J. Shen, “Deep visual attention prediction,”IEEE Trans- actions on Image Processing, vol. 27, no. 5, pp. 2368–2378, 2018

    2018

  51. [59]

    A review of approaches for predicting drug–drug interactions based on machine learning,

    K. Han, P. Cao, Y . Wang, F. Xie, J. Ma, M. Yu, J. Wang, Y . Xu, Y . Zhang, and J. Wan, “A review of approaches for predicting drug–drug interactions based on machine learning,”Frontiers in Pharmacology, vol. 12, 2022

    2022

  52. [60]

    Training-free object- background compositional t2i via dynamic spatial guidance and multi- path pruning,

    Y . Deng, D. Mould, P. L. Rosin, and Y .-K. Lai, “Training-free object- background compositional t2i via dynamic spatial guidance and multi- path pruning,”arXiv preprint arXiv:2604.09850, 2026

    2026 arXiv

  53. [61]

    Sinkhorn distances: Lightspeed computation of optimal transport,

    M. Cuturi, “Sinkhorn distances: Lightspeed computation of optimal transport,”Advances in neural information processing systems, vol. 26, 2013

    2013

  54. [62]

    Optimal transport for single-cell and spatial omics,

    C. Bunne, G. Schiebinger, A. Krause, A. Regev, and M. Cuturi, “Optimal transport for single-cell and spatial omics,”Nature Reviews Methods Primers, vol. 4, no. 1, p. 58, 2024

    2024

  55. [63]

    Adapting large language models by integrating collaborative semantics for recommendation,

    B. Zheng, Y . Hou, H. Lu, Y . Chen, W. X. Zhao, M. Chen, and J.-R. Wen, “Adapting large language models by integrating collaborative semantics for recommendation,” in2024 IEEE 40th International Conference on Data Engineering (ICDE). IEEE, 2024, pp. 1435–1448

    2024

  56. [64]

    Clip is also an efficient segmenter: A text-driven approach for weakly supervised semantic segmentation,

    Y . Lin, M. Chen, W. Wang, B. Wu, K. Li, B. Lin, H. Liu, and X. He, “Clip is also an efficient segmenter: A text-driven approach for weakly supervised semantic segmentation,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2023, pp. 15 305–15 314

    2023

  57. [65]

    Hierarchical dense correlation distillation for few-shot segmentation,

    B. Peng, Z. Tian, X. Wu, C. Wang, S. Liu, J. Su, and J. Jia, “Hierarchical dense correlation distillation for few-shot segmentation,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recogni- tion, 2023, pp. 23 641–23 651

    2023

  58. [66]

    Pfenet++: Boosting few-shot semantic segmentation with the noise-filtered context- aware prior mask,

    X. Luo, Z. Tian, T. Zhang, B. Yu, Y . Y . Tang, and J. Jia, “Pfenet++: Boosting few-shot semantic segmentation with the noise-filtered context- aware prior mask,”IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 46, no. 2, pp. 1273–1289, 2024

    2024

  59. [67]

    Drnet: Disentanglement and recombination network for few-shot semantic segmentation,

    Z. Chang, X. Gao, N. Li, H. Zhou, and Y . Lu, “Drnet: Disentanglement and recombination network for few-shot semantic segmentation,”IEEE Transactions on Circuits and Systems for Video Technology, vol. 34, no. 7, pp. 5560–5574, 2024

    2024

  60. [68]

    Rpmg- fss: Robust prior mask guided few-shot semantic segmentation,

    L. Zhang, X. Zhang, Q. Wang, W. Wu, X. Chang, and J. Liu, “Rpmg- fss: Robust prior mask guided few-shot semantic segmentation,”IEEE Transactions on Circuits and Systems for Video Technology, vol. 33, no. 11, pp. 6609–6621, 2023

    2023

  61. [69]

    Vrp-sam: Sam with visual reference prompt,

    Y . Sun, J. Chen, S. Zhang, X. Zhang, Q. Chen, G. Zhang, E. Ding, J. Wang, and Z. Li, “Vrp-sam: Sam with visual reference prompt,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2024, pp. 23 565–23 574

    2024

  62. [70]

    Weakly supervised few-shot semantic segmentation via pseudo mask enhancement and meta learning,

    M. Zhang, Y . Zhou, B. Liu, J. Zhao, R. Yao, Z. Shao, and H. Zhu, “Weakly supervised few-shot semantic segmentation via pseudo mask enhancement and meta learning,”IEEE Transactions on Multimedia, vol. 25, pp. 7980–7991, 2023. Jiaqi Mareceived the B.S. degree in Harbin Univer- ...

    2023

  63. [2013]

    His research interests include big media analysis and computer vision

    He is currently a Professor at the Nanjing University of Science and Technology. His research interests include big media analysis and computer vision. He received the Best Paper Award at ACM Multimedia Asia 2020 and 2024. He is currently serving as an associate editor for IEE...

    2020

  64. [2016]

    His research interests include computer vision and machine learning

    He is an Associate Editor of IEEE T-IP and Pattern Recognition Journals, and Area Chairs of several international conferences, such as ICLR. His research interests include computer vision and machine learning. Fang Zhaoreceived the Ph.D. degree from the Na- tional Laboratory o...

    2015

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.