Pith. sign in

REVIEW 4 major objections 5 minor 1 cited by

Overcoming Support Dilution for Robust Few-shot Semantic Segmentation

T0 review · 4 major / 5 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read Support dilution in few-shot segmentation can be diagnosed with a contribution index and fixed with symmetric correlation and pruning.

desk verdict Solid study of a real phenomenon, with a load-bearing derivation gap in the pruning step; worth reviewing, not worth accepting as-is. read the letter →

arxiv 2501.13529 v1 pith:N3R25FEY submitted 2025-01-23 cs.CV cs.LG

Wailing Tang , Biqi Yang , Pheng-Ann Heng , Yun-Hui Liu , Chi-Wing Fu This is my paper · ORCID
classification cs.CVcs.LG
keywords few-shotsemanticsegmentationsupportdilutioncontributionindexsymmetriccorrelationimagepruningcross-attentionmulti-shot
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 establish that a common failure mode in few-shot semantic segmentation, called support dilution, can be diagnosed and fixed. Support dilution is what happens when adding more example (support) images stops improving, and then starts hurting, segmentation of the query image: the useful supports get drowned out by irrelevant ones. The proposed remedy has three parts: a contribution index that scores each support by its attention to the query, a Symmetric Correlation module that makes the best-matching support dominate the attention, and a pruning operation that discards low-scoring supports before correlation is computed. If the paper is right, an FSS system can take a large, noisy support pool and keep getting better masks as the pool grows, which is exactly what real-world use on web-collected images would require.

What carries the argument

The load-bearing object is the contribution index $\delta(x_s)$, defined as the average over a support's tokens of their maximum attention weight to any query token; it is meant to quantify how much a support truly helps a query. Symmetric Correlation rewrites the attention score so that keys and queries share one function $f(x)=f_1(x)\frac{f_2(x)}{\|f_2(x)\|_2}$, with a magnitude part for objectness and a normalized angle part for similarity; symmetry makes the score maximal if and only if support and query are identical. Support Image Pruning turns subset selection into a greedy maximization of a Jensen-derived lower bound on the summed contribution index, reducing cost from enumerating all subsets to $O(N'\times N)$. These three pieces work together: the index measures dilution, the symmetric attention resists it, and pruning removes the worst diluters before attention is applied.

What would settle it

Compute an oracle contribution for each support by removing it, re-running segmentation, and measuring the mIoU change on a fixed set of queries. If the paper's contribution index has low rank correlation with that oracle, or if mIoU stops climbing when support pools are deliberately contaminated with many irrelevant images, the central claim is refuted.

Watch

Extended reading notes

Core claim

The central claim is that support dilution is a real and measurable effect, and that it can be overcome by coupling support scoring with a symmetric attention design. The paper shows that in existing methods the deviation between the contribution index of a perfect "upper-bound" support and the average of the other supports shrinks as the number of supports grows, and that segmentation mIoU falls with it. Its Symmetric Correlation uses the same network to produce keys and queries plus a normalization into magnitude and angle parts, so that attention is maximized exactly when support and query features match; this keeps high-contributed supports standing out even in a noisy pool. Support Image Pruning then greedily keeps the supports that maximize a lower bound on total contribution, which the paper justifies via Jensen's inequality, and the full pipeline reports mIoU that keeps climbing from 1 to 70 shots on both COCO-20i and PASCAL-5i across three backbones.

Load-bearing premise

The load-bearing premise is that the contribution index, which is computed from attention weights alone, correctly ranks how much each support truly helps segment the query; if that ranking is wrong, pruning will discard the useful supports and Symmetric Correlation will amplify the wrong ones.

Editorial extensions

If this is right

  • Segmentation quality continues to rise as the support count grows from 1 to 70 shots, instead of flattening or dropping.
  • The Symmetric Correlation and pruning components can be inserted into prior FSS methods and improve their multi-shot results, suggesting the fix is architectural rather than benchmark-specific.
  • When the support pool exceeds 30 images, pruning to 30 supports keeps the gains while holding inference time and memory roughly constant.
  • The method transfers across domain gaps and tolerates automatically generated support masks, so web-searched supports can be used without manual filtering.

Reading between the lines

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

  • The paper's diagnosis suggests a general stress test for any attention-based few-shot learner: record how the contribution deviation changes as the support pool grows; a falling deviation predicts vulnerability to dilution even before mIoU drops.
  • Because the contribution index is defined purely from attention weights, an obvious extension is to validate it against an oracle contribution measured by leave-one-out changes in mask quality; the pruning objective could then be replaced or calibrated by that oracle.
  • The same dilution mechanism likely applies to other tasks where many weakly relevant examples are pooled, such as few-shot object detection or retrieval-augmented generation, wherever soft attention aggregates a large noisy context.
  • The Jensen lower bound used for pruning drops the softmax partition function, so the pruning score is unnormalized; a testable variant would divide by the partition or normalize by feature magnitude to see whether pruning choices change.
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

4 major / 5 minor

Summary. This paper addresses the problem of "support dilution" in few-shot semantic segmentation (FSS), where increasing the number of support images can degrade rather than improve segmentation quality. The authors propose three components: a contribution index δ (Eq. 4) based on maximum softmax attention weights; a Symmetric Correlation (SC) module that uses a shared normalization f(x)=f1(x)f2(x)/||f2(x)|| for keys and queries (Eqs. 5-6); and Support Image Pruning, a greedy subset-selection algorithm based on an approximate objective θ (Eq. 9) derived from δ. Experiments on COCO-20i and PASCAL-5i with ResNet-50, ResNet-101, and Swin-B backbones show consistent mIoU improvements over DCAMA and MSANet as N grows from 1 to 70, together with cross-domain, plug-and-play, online, and real-world demonstrations.

Significance. The support-dilution problem is practically important and has received little attention, and the paper gives a clear motivation with extensive experiments. If the pruning derivation can be made correct or empirically validated, the method would be a valuable contribution: the plug-in gains over HDMNet, MSANet, and DCAMA and the reduced inference cost at large N are concrete and useful. The manuscript is generally well written and transparent about its training setup, and the online and real-world experiments strengthen the practical story. However, the theoretical justification of the pruning objective is currently not sound, and the contribution-index validation is incomplete, so the central claims cannot be fully accepted without revision.

major comments (4)
  1. [IV-E, Eq. (8)] Equation (8) is not a valid Jensen bound as stated. The left-hand side is an average over support tokens of a maximum softmax attention probability, while the right-hand side is a raw dot product of f applied to averaged tokens. The softmax partition function and the max operation are removed, and no assumptions on the range or distribution of the logits are given that would make the inequality hold. Because Eq. (9) and Algorithm 1 replace the stated objective in Eq. (7) with this lower bound, the pruning stage currently optimizes a different and unvalidated surrogate rather than the contribution index it claims to maximize. Moreover, δ in Eq. (4) depends on the whole support set through the softmax denominator, so the objective in Eq. (7) is not a simple sum of per-support independent terms; the greedy algorithm does not solve Eq. (7). Please provide a correct derivation with explicit assumptions, or recharacterize Eq. (9) as an empirical heuristic and validate it directly against δ and against segmentation performance, for example by comparing selected subsets with those chosen by an oracle.
  2. [IV-D] The claim that using the same network f for keys and queries "guarantees the attention function to have an unique maximum point" for identical support-query pairs is not proved and does not follow from symmetry alone. A symmetric score f(x)^T f(y) can have f(x)^T f(z) >= f(x)^T f(x) for some z different from x unless f is sufficiently injective and the magnitude term f1 is positive; the normalization f(x)=f1(x)f2(x)/||f2(x)|| does not by itself ensure this. Furthermore, the upper-bound experiments in Fig. 4 and Table I partly test a property that SC is explicitly designed to enforce, so those results should be interpreted as a sanity check of the architecture rather than as independent evidence that high-contributed supports are being recognized. The benchmark improvements in Fig. 7 remain the primary evidence for the method.
  3. [IV-C, Table VII] The contribution index δ is defined purely from attention weights and is never checked against any ground-truth notion of support contribution. Table VII compares Support Image Pruning with feature-distance retrieval baselines but does not include an oracle baseline (for example, selecting supports by measured segmentation gain or by the known upper-bound support) or a random-selection baseline, so the experiments cannot show whether the pruned subset actually contains the true high-contributed supports. Adding such baselines, or an ablation in which the contribution ranking is corrupted or permuted, would directly test the validity of δ and of the pruning objective.
  4. [V-C, Fig. 7] All benchmark numbers are reported as single point estimates without error bars or multiple seeds. Some of the claimed gains at small N (for example, 1-shot in several folds of Fig. 7) are small and could be within run-to-run variation; given the headline claim that mIoU "keeps climbing from 1 to 70 shots", please report means and standard deviations over at least three random seeds for the main comparisons, or state the number of runs used and provide a variance analysis.
minor comments (5)
  1. [V-B] The word "backpropogation" should be "backpropagation".
  2. [Fig. 8 caption] The caption contains the typo "Transfromer"; it should be "Transformer".
  3. [V-E] The section text uses "Plug-and-paly"; this should be "Plug-and-play".
  4. [References] Reference [12] is cited in the text as the source of ResNet, but the bibliography entry is for Hariharan et al.; please update the reference to the original ResNet paper.
  5. [Eqs. (3), (6), (12)] The softmax expressions contain an extra closing parenthesis after the argument; please clean up the notation for clarity.

Circularity Check

1 steps flagged · score 2.0 of 10

Minor self-definitional upper-bound illustration; central benchmark claims are externally grounded.

  1. self definitional [Sec. I (Symmetric Correlation) and Sec. IV-D, Eqs. 5-6, Fig. 4]
    "The key idea is to ensure the correlation score to reach the maximum when and only when we input an identical support-query pair. With this constraint, our upper-bound support can permanently obtain the largest contribution value, and other high-contributed support features are simultaneously consolidated depending on their visual similarity with the query."

    The contribution index δ (Eq. 4) is defined as the mean over support tokens of the maximum softmax attention weight. Symmetric Correlation (Eqs. 5-6) is explicitly constructed with a shared f and normalized features so that an identical support-query pair receives maximal attention. Hence the observation that the upper-bound support has the largest δ and a high deviation Δ in Fig. 4 is not an independent empirical finding but an algebraic consequence of the definitions. The paper presents this as evidence that SC 'preserves and enhances the high-contributed support features,' yet the metric used to demonstrate preservation is the same attention quantity the module was designed to maximize.

full rationale

The paper's main claims—mIoU gains on COCO-20i and PASCAL-5i, plug-and-play improvements on HDMNet/MSANet/DCAMA, and cross-domain results—are all evaluated against external segmentation masks and prior methods, so they do not reduce to the paper's own definitions. The only notable circular step is the upper-bound-support demonstration in Sec. IV-D: the contribution index δ is mean max attention, and SC is architected so that an identical support-query pair gets maximal attention; therefore the large deviation for SC in Fig. 4 largely confirms a property built into the design rather than providing independent evidence. This is motivational rather than load-bearing for the headline benchmark results. Separately, the Jensen lower bound in Eq. 8 is asserted without a valid derivation—the softmax partition function and max operation are dropped—but this is an unverified mathematical step and a correctness risk, not a circularity; the pruning module could fail to maximize the stated δ even if it maximizes θ. No self-citation chain, uniqueness import, or benchmark-fitted prediction is present. Accordingly, the circularity score is low.

Assumptions & free parameters 2 free parameters · 4 assumptions · 1 invented entities

The central derivation rests on a hand-picked pruning budget and several unproven assumptions about the attention-based contribution index and the symmetric-normalized correlation. The contribution index has no external labels, so its fidelity as a measure of true support utility is an assumption rather than a demonstrated fact.

free parameters (2)
  • N' (pruning budget) = 30
    When N > 30, the method prunes to N'=30 fixed supports (Sec. V-B, Tab. VI). This threshold is hand-chosen; no sensitivity analysis across N' is reported.
  • f2 initialization = weights zeros, biases ones
    SC modules initialize f2 weights to 0 and biases to 1, assuming identical objectness for all supports (Sec. V-B). This initialization choice affects learned attention but is not swept.
assumptions (4)
  • domain assumption Support images can be ranked by visual similarity to the query, and high-contributed supports are the visually similar ones.
    Sec. IV-C defines contribution through attention weights; no ground-truth support contribution labels are used, so the ranking's validity is assumed.
  • ad hoc to paper The learned function f makes attention reach a unique maximum exactly when support and query features are identical.
    Eqs. 5-6 assert the upper-bound constraint; injectivity and uniqueness are not proven for the trained f.
  • ad hoc to paper Jensen's inequality justifies replacing per-token softmax attention with an unnormalized average-token lower bound.
    Sec. IV-E Eqs. 8-9; the softmax partition function is dropped and f is nonlinear, so the inequality is not established as written.
  • domain assumption Features learned on COCO transfer to PASCAL-5i, SUNRGBD, KITTI, and web images without adaptation.
    Cross-domain and real-world experiments (Sec. V-D, V-H) rely on this transferability; no domain adaptation is used.
invented entities (1)
  • Contribution index δ(x_s)
    purpose: Quantitative score for each support's contribution to the query; drives pruning and is used to measure dilution.
    Defined from attention weights in Eq. 4; no external labels or benchmark establish that δ correlates with true segmentation benefit.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Overcoming Support Dilution for Robust Few-shot Semantic Segmentation." pith.science (2026). https://pith.science/paper/N3R25FEY

@misc{pith2026250113529,
  author       = {Pith},
  title        = {Pith review of: Overcoming Support Dilution for Robust Few-shot Semantic Segmentation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/N3R25FEY}},
  note         = {Machine review of arXiv:2501.13529}
}
read the original abstract

Few-shot Semantic Segmentation (FSS) is a challenging task that utilizes limited support images to segment associated unseen objects in query images. However, recent FSS methods are observed to perform worse, when enlarging the number of shots. As the support set enlarges, existing FSS networks struggle to concentrate on the high-contributed supports and could easily be overwhelmed by the low-contributed supports that could severely impair the mask predictions. In this work, we study this challenging issue, called support dilution, our goal is to recognize, select, preserve, and enhance those high-contributed supports in the raw support pool. Technically, our method contains three novel parts. First, we propose a contribution index, to quantitatively estimate if a high-contributed support dilutes. Second, we develop the Symmetric Correlation (SC) module to preserve and enhance the high-contributed support features, minimizing the distraction by the low-contributed features. Third, we design the Support Image Pruning operation, to retrieve a compact and high quality subset by discarding low-contributed supports. We conduct extensive experiments on two FSS benchmarks, COCO-20i and PASCAL-5i, the segmentation results demonstrate the compelling performance of our solution over state-of-the-art FSS approaches. Besides, we apply our solution for online segmentation and real-world segmentation, convincing segmentation results showing the practical ability of our work for real-world demonstrations.

Figures

Figures reproduced from arXiv: 2501.13529 by the authors.

Figure 1
Figure 1. When the number of supports gets larger, SOTA [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Left: the performance of DCAMA cannot gain consistent improvements when the number of shots [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. The supports in the same category can have significant [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (11 more)
Figure 4
Figure 4. Figure 4: Left: The deviation value ∆ v.s. the number of shots N. Right: segmentation mIoU v.s. the number of shots N. Experiments on COCO-20i Fold 1 under the upper-bound setting. one should result in a higher contribution value with Eq. 4, if not, the relative deviation can re…
Figure 5
Figure 5. Figure 5: The pipeline of our network. We introduce the flow [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]
Figure 6
Figure 6. Figure 6: The dataflow and intermediate feature shapes of the [PITH_FULL_IMAGE:figures/full_fig_p008_6.png]
Figure 7
Figure 7. Figure 7: We compare the segmentation mIoU metric with MSANet and DCAMA. We report results on two FSS benchmarks, [PITH_FULL_IMAGE:figures/full_fig_p009_7.png]
Figure 8
Figure 8. Figure 8: Qualitative comparison with DCAMA on COCO-20 [PITH_FULL_IMAGE:figures/full_fig_p011_8.png]
Figure 9
Figure 9. Figure 9: Typical results on PASCAL-5i . Note that the training support-query pairs are from COCO-20i while the testing support￾query pairs are from PASCAL-5i , N = 5, backbone is Swin-Transformer-Base. Thanks to Symmetric Correlation (SC), our method is robust to data distribut…
Figure 11
Figure 11. Figure 11: Mask results of using noisy support masks generated [PITH_FULL_IMAGE:figures/full_fig_p012_11.png]
Figure 12
Figure 12. Figure 12: Qualitative results on indoor dataset SUNRGBD (top two rows) and autonomous driving dataset KITTI (bottom two [PITH_FULL_IMAGE:figures/full_fig_p013_12.png]
Figure 13
Figure 13. Figure 13: Qualitative results for complex (i.e., multi-category) [PITH_FULL_IMAGE:figures/full_fig_p013_13.png]
Figure 14
Figure 14. Figure 14: Real-time qualitative results on queries that we [PITH_FULL_IMAGE:figures/full_fig_p014_14.png]
Figure 15
Figure 15. Figure 15: Typical failure cases in 1-shot testing. Left: the model [PITH_FULL_IMAGE:figures/full_fig_p014_15.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

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

  1. Toward Robust In-Context Segmentation via Concept Guidance

    cs.CV 2026-06 unverdicted novelty 6.0 of 10

    CG-ICS uses an MLLM to generate candidate textual concepts, scores them with SAM3 in a tree search, and combines the best concept with a visual exemplar to segment the query stably.

Reference graph

Works this paper leans on

58 extracted references · 49 canonical work pages · cited by 1 Pith paper

  1. [1]

    Avrithis, Y ., Tolias, G.: Hough pyramid matching: Speeded-up geometry re-ranking for large scale image retrieval. Int. J. Comp. Vision 107, 1–19 (2014)

    work page 2014

  2. [2]

    Chen, B.k., Gong, C., Yang, J.: Importance-aware semantic segmentation for autonomous driving system. pp. 1504–1510 (2017)

    work page 2017

  3. [3]

    In: AAAI Conf

    Chen, Z., Fu, Y ., Chen, K., Jiang, Y .G.: Image block augmentation for one-shot learning. In: AAAI Conf. on Artificial Intell. (AAAI). vol. 33, pp. 3379–3386 (2019)

    work page 2019

  4. [4]

    In: IEEE/CVF Conf

    Chen, Z., Fu, Y ., Wang, Y .X., Ma, L., Liu, W., Hebert, M.: Image deformation meta-networks for one-shot learning. In: IEEE/CVF Conf. on Computer Vision and Pattern Recognition (CVPR). pp. 8680–8689 (2019)

    work page 2019

  5. [5]

    In: IEEE/CVF Conf

    Dalal, N., Triggs, B.: Histograms of oriented gradients for human detec- tion. In: IEEE/CVF Conf. on Computer Vision and Pattern Recognition (CVPR). vol. 1, pp. 886–893 (2005)

    work page 2005

  6. [6]

    Dong, N., Xing, E.P.: Few-shot semantic segmentation with prototype learning. vol. 3, p. 4 (2018)

    work page 2018

  7. [7]

    In: IEEE/CVF Int

    Feng, C.M., Yu, K., Liu, Y ., Khan, S., Zuo, W.: Diverse data augmenta- tion with diffusions for effective test-time prompt tuning. In: IEEE/CVF Int. Conf. on Computer Vision (ICCV). pp. 2704–2714 (2023)

    work page 2023

  8. [8]

    Finn, C., Abbeel, P., Levine, S.: Model-agnostic meta-learning for fast adaptation of deep networks. In: Int. Conf. on Machine Learning (ICML). pp. 1126–1135 (2017)

    work page 2017

Show all 58 references
  1. [9]

    In: IEEE/CVF Conf

    Geiger, A., Lenz, P., Urtasun, R.: Are we ready for autonomous driving? the kitti vision benchmark suite. In: IEEE/CVF Conf. on Computer Vision and Pattern Recognition (CVPR). pp. 3354–3361 (2012)

    2012

  2. [10]

    In: IEEE/CVF Conf

    Ghiasi, G., Cui, Y ., Srinivas, A., Qian, R., Lin, T.Y ., Cubuk, E.D., Le, Q.V ., Zoph, B.: Simple copy-paste is a strong data augmentation method for instance segmentation. In: IEEE/CVF Conf. on Computer Vision and Pattern Recognition (CVPR). pp. 2918–2928 (2021)

    2021

  3. [11]

    Journal of the Royal Statistical Society: Series B (Methodological) 14(1), 107–114 (1952)

    Good, I.J.: Rational decisions. Journal of the Royal Statistical Society: Series B (Methodological) 14(1), 107–114 (1952)

    1952

  4. [12]

    In: European Conf

    Hariharan, B., Arbel ´aez, P., Girshick, R., Malik, J.: Simultaneous detection and segmentation. In: European Conf. on Computer Vision (ECCV). pp. 297–312 (2014)

    2014

  5. [13]

    In: AAAI Conf

    Hu, T., Yang, P., Zhang, C., Yu, G., Mu, Y ., Snoek, C.G.: Attention-based multi-context guiding for few-shot semantic segmentation. In: AAAI Conf. on Artificial Intell. (AAAI). vol. 33, pp. 8441–8448 (2019)

    2019

  6. [14]

    IEEE Jour

    Huang, B., Tian, J., Zhang, H., Luo, Z., Qin, J., Huang, C., He, X., Luo, Y ., Zhou, Y ., Dan, G., et al.: Deep semantic segmentation feature-based radiomics for the classification tasks in medical image analysis. IEEE Jour. of Biomedical and Health Informatics 25(7), 2655–2664 (2020)

    2020

  7. [15]

    In: IEEE/CVF Int

    Huang, Z., Wang, X., Huang, L., Huang, C., Wei, Y ., Liu, W.: Ccnet: Criss-cross attention for semantic segmentation. In: IEEE/CVF Int. Conf. on Computer Vision (ICCV). pp. 603–612 (2019)

    2019

  8. [16]

    arXiv preprint arXiv:2206.09667 (2022)

    Iqbal, E., Safarov, S., Bang, S.: Msanet: Multi-similarity and at- tention guidance for boosting few-shot segmentation. arXiv preprint arXiv:2206.09667 (2022)

    2022 arXiv

  9. [17]

    In: IEEE/CVF Conf

    Jamal, M.A., Qi, G.J.: Task agnostic meta-learning for few-shot learning. In: IEEE/CVF Conf. on Computer Vision and Pattern Recognition (CVPR). pp. 11719–11727 (2019)

    2019

  10. [18]

    IEEE Access 9, 83002–83024 (2021)

    Khan, M.Z., Gajendran, M.K., Lee, Y ., Khan, M.A.: Deep neural architectures for medical image semantic segmentation. IEEE Access 9, 83002–83024 (2021)

    2021

  11. [19]

    IEEE Trans

    Lang, C., Cheng, G., Tu, B., Li, C., Han, J.: Base and meta: A new perspective on few-shot segmentation. IEEE Trans. Pattern Anal. & Mach. Intell. (T-PAMI) (2023)

    2023

  12. [20]

    In: IEEE/CVF Conf

    Li, G., Jampani, V ., Sevilla-Lara, L., Sun, D., Kim, J., Kim, J.: Adap- tive prototype learning and allocation for few-shot segmentation. In: IEEE/CVF Conf. on Computer Vision and Pattern Recognition (CVPR). pp. 8334–8343 (2021)

    2021

  13. [21]

    In: IEEE/CVF Conf

    Li, H., Eigen, D., Dodge, S., Zeiler, M., Wang, X.: Finding task-relevant features for few-shot learning by category traversal. In: IEEE/CVF Conf. on Computer Vision and Pattern Recognition (CVPR). pp. 1–10 (2019)

    2019

  14. [22]

    IEEE Sensors Jour

    Lin, C.M., Tsai, C.Y ., Lai, Y .C., Li, S.A., Wong, C.C.: Visual object recognition and pose estimation based on a deep semantic segmentation network. IEEE Sensors Jour. 18(22), 9370–9381 (2018)

    2018

  15. [23]

    In: IEEE/CVF Conf

    Lin, G., Milan, A., Shen, C., Reid, I.: Refinenet: Multi-path refinement networks for high-resolution semantic segmentation. In: IEEE/CVF Conf. on Computer Vision and Pattern Recognition (CVPR). pp. 1925– 1934 (2017)

    2017

  16. [24]

    In: IEEE/CVF Int

    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: IEEE/CVF Int. Conf. on Computer Vision (ICCV). pp. 10012–10022 (2021) 15

    2021

  17. [25]

    Lowe, D.G.: Distinctive image features from scale-invariant keypoints. Int. J. Comp. Vision 60, 91–110 (2004)

    2004

  18. [26]

    In: IEEE/CVF Int

    Min, J., Kang, D., Cho, M.: Hypercorrelation squeeze for few-shot segmentation. In: IEEE/CVF Int. Conf. on Computer Vision (ICCV). pp. 6941–6952 (2021)

    2021

  19. [27]

    In: IEEE/CVF Int

    Nguyen, K., Todorovic, S.: Feature weighting and boosting for few-shot segmentation. In: IEEE/CVF Int. Conf. on Computer Vision (ICCV). pp. 622–631 (2019)

    2019

  20. [28]

    arXiv preprint arXiv:2304.07193 (2023)

    Oquab, M., Darcet, T., Moutakanni, T., V o, H., Szafraniec, M., Khalidov, V ., Fernandez, P., Haziza, D., Massa, F., El-Nouby, A., et al.: Dinov2: Learning robust visual features without supervision. arXiv preprint arXiv:2304.07193 (2023)

    2023 arXiv

  21. [29]

    Paszke, A., Gross, S., Massa, F., Lerer, A., Bradbury, J., Chanan, G., Killeen, T., Lin, Z., Gimelshein, N., Antiga, L., et al.: Pytorch: An imperative style, high-performance deep learning library. Conf. and Workshop on Neural Information Processing Systems (NeurIPS) 32 (2019)

    2019

  22. [30]

    In: IEEE/CVF Conf

    Peng, B., Tian, Z., Wu, X., Wang, C., Liu, S., Su, J., Jia, J.: Hierarchical dense correlation distillation for few-shot segmentation. In: IEEE/CVF Conf. on Computer Vision and Pattern Recognition (CVPR). pp. 23641– 23651 (2023)

    2023

  23. [31]

    Radford, A., Kim, J.W., Hallacy, C., Ramesh, A., Goh, G., Agarwal, S., Sastry, G., Askell, A., Mishkin, P., Clark, J., et al.: Learning transferable visual models from natural language supervision. In: Int. Conf. on Machine Learning (ICML). pp. 8748–8763 (2021)

    2021

  24. [32]

    Ravi, S., Larochelle, H.: Optimization as a model for few-shot learning. In: Int. Conf. on Learning Representations (ICLR) (2016)

    2016

  25. [33]

    arxiv 2024

    Ren, T., Liu, S., Zeng, A., Lin, J., Li, K., Cao, H., Chen, J., Huang, X., Chen, Y ., Yan, F., et al.: Grounded sam: Assembling open-world models for diverse visual tasks. arxiv 2024. arXiv preprint arXiv:2401.14159

    2024 arXiv

  26. [34]

    Ronneberger, O., Fischer, P., Brox, T.: U-net: Convolutional networks for biomedical image segmentation. In: Int. Conf. on MICCAI. pp. 234–

  27. [35]

    Schwarz, M., Milan, A., Periyasamy, A.S., Behnke, S.: Rgb-d object detection and semantic segmentation for autonomous manipulation in clutter. Int. Jour. Robotics Research (IJRR) 37(4-5), 437–451 (2018)

    2018

  28. [36]

    arXiv preprint arXiv:1709.03410 (2017)

    Shaban, A., Bansal, S., Liu, Z., Essa, I., Boots, B.: One-shot learning for semantic segmentation. arXiv preprint arXiv:1709.03410 (2017)

    2017 arXiv

  29. [37]

    In: European Conf

    Shi, X., Wei, D., Zhang, Y ., Lu, D., Ning, M., Chen, J., Ma, K., Zheng, Y .: Dense cross-query-and-support attention weighted mask aggregation for few-shot segmentation. In: European Conf. on Computer Vision (ECCV). pp. 151–168 (2022)

    2022

  30. [38]

    arXiv preprint arXiv:1409.1556 (2014)

    Simonyan, K., Zisserman, A.: Very deep convolutional networks for large-scale image recognition. arXiv preprint arXiv:1409.1556 (2014)

    2014 arXiv

  31. [39]

    Snell, J., Swersky, K., Zemel, R.: Prototypical networks for few- shot learning. Conf. and Workshop on Neural Information Processing Systems (NeurIPS) 30 (2017)

    2017

  32. [40]

    In: IEEE/CVF Conf

    Song, S., Lichtenberg, S.P., Xiao, J.: Sun rgb-d: A rgb-d scene under- standing benchmark suite. In: IEEE/CVF Conf. on Computer Vision and Pattern Recognition (CVPR). pp. 567–576 (2015)

    2015

  33. [41]

    In: IEEE/CVF Conf

    Sung, F., Yang, Y ., Zhang, L., Xiang, T., Torr, P.H., Hospedales, T.M.: Learning to compare: Relation network for few-shot learning. In: IEEE/CVF Conf. on Computer Vision and Pattern Recognition (CVPR). pp. 1199–1208 (2018)

    2018

  34. [42]

    IEEE Trans

    Tian, Z., Zhao, H., Shu, M., Yang, Z., Li, R., Jia, J.: Prior guided feature enrichment network for few-shot segmentation. IEEE Trans. Pattern Anal. & Mach. Intell. (T-PAMI) 44(2), 1050–1065 (2020)

    2020

  35. [43]

    arXiv preprint arXiv:2302.07944 (2023)

    Trabucco, B., Doherty, K., Gurinas, M., Salakhutdinov, R.: Effective data augmentation with diffusion models. arXiv preprint arXiv:2302.07944 (2023)

    2023 arXiv

  36. [44]

    Vinyals, O., Blundell, C., Lillicrap, T., Wierstra, D., et al.: Matching net- works for one shot learning. Conf. and Workshop on Neural Information Processing Systems (NeurIPS) 29 (2016)

    2016

  37. [45]

    https://doi.org/10.5281/zenodo.5711226, https://github.com/wkentaro/ labelme

    Wada, K.: Labelme: Image Polygonal Annotation with Python. https://doi.org/10.5281/zenodo.5711226, https://github.com/wkentaro/ labelme

    doi

  38. [46]

    In: IEEE/CVF Int

    Wang, K., Liew, J.H., Zou, Y ., Zhou, D., Feng, J.: Panet: Few-shot image semantic segmentation with prototype alignment. In: IEEE/CVF Int. Conf. on Computer Vision (ICCV). pp. 9197–9206 (2019)

    2019

  39. [47]

    arxiv 2020

    Wang, X., Huang, T., Darrell, T., Gonzalez, J., Yu, F.: Frustrat- ingly simple few-shot object detection. arxiv 2020. arXiv preprint arXiv:2003.06957

    2020 arXiv

  40. [48]

    Sensors 21(7), 2280 (2021)

    Wong, C.C., Yeh, L.Y ., Liu, C.C., Tsai, C.Y ., Aoyama, H.: Manipula- tion planning for object re-orientation based on semantic segmentation keypoint detection. Sensors 21(7), 2280 (2021)

    2021

  41. [49]

    In: IEEE/CVF Int

    Wu, Z., Shi, X., Lin, G., Cai, J.: Learning meta-class memory for few-shot semantic segmentation. In: IEEE/CVF Int. Conf. on Computer Vision (ICCV). pp. 517–526 (2021)

    2021

  42. [50]

    Neural Networks 157, 460–470 (2023)

    Xiao, X., Zhao, Y ., Zhang, F., Luo, B., Yu, L., Chen, B., Yang, C.: Baseg: Boundary aware semantic segmentation for autonomous driving. Neural Networks 157, 460–470 (2023)

    2023

  43. [51]

    In: IEEE/CVF Int

    Xu, Q., Zhao, W., Lin, G., Long, C.: Self-calibrated cross attention net- work for few-shot segmentation. In: IEEE/CVF Int. Conf. on Computer Vision (ICCV). pp. 655–665 (2023)

    2023

  44. [52]

    In: IEEE/CVF Conf

    Yang, M., Yu, K., Zhang, C., Li, Z., Yang, K.: Denseaspp for semantic segmentation in street scenes. In: IEEE/CVF Conf. on Computer Vision and Pattern Recognition (CVPR). pp. 3684–3692 (2018)

    2018

  45. [53]

    Frontiers in Oncology 11, 638182 (2021)

    Yang, R., Yu, Y .: Artificial convolutional neural network in object detection and semantic segmentation for medical imaging analysis. Frontiers in Oncology 11, 638182 (2021)

    2021

  46. [54]

    arXiv preprint arXiv:1511.07122 (2015)

    Yu, F., Koltun, V .: Multi-scale context aggregation by dilated convolu- tions. arXiv preprint arXiv:1511.07122 (2015)

    2015 arXiv

  47. [55]

    In: IEEE/CVF Conf

    Zhang, C., Lin, G., Liu, F., Yao, R., Shen, C.: Canet: Class-agnostic segmentation networks with iterative refinement and attentive few- shot learning. In: IEEE/CVF Conf. on Computer Vision and Pattern Recognition (CVPR). pp. 5217–5226 (2019)

    2019

  48. [56]

    Zhang, G., Kang, G., Yang, Y ., Wei, Y .: Few-shot segmentation via cycle-consistent transformer. Conf. and Workshop on Neural Information Processing Systems (NeurIPS) 34, 21984–21996 (2021)

    2021

  49. [57]

    IEEE Trans

    Zhang, X., Wei, Y ., Yang, Y ., Huang, T.S.: Sg-one: Similarity guidance network for one-shot semantic segmentation. IEEE Trans. on Cybernet- ics 50(9), 3855–3865 (2020)

    2020

  50. [58]

    In: IEEE/CVF Conf

    Zhao, H., Shi, J., Qi, X., Wang, X., Jia, J.: Pyramid scene parsing net- work. In: IEEE/CVF Conf. on Computer Vision and Pattern Recognition (CVPR). pp. 2881–2890 (2017)

    2017

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.