Pith. sign in

REVIEW 4 major objections 6 minor 32 references

B2Net: Camouflaged Object Detection via Boundary Aware and Boundary Fusion

T0 review · 4 major / 6 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read B2Net claims that running a boundary-aware module twice, with a cross-scale fusion cascade in between, produces sharper camouflaged-object boundaries and outperforms 15 published methods on three standard benchmarks.

desk verdict A solid incremental COD paper with a defensible architecture, but the reported margins over FSPNet are thin and the missing code/seeds make the headline claim not yet independently verifiable. read the letter →

arxiv 2501.00426 v1 pith:QKXQN7IL submitted 2024-12-31 cs.CV cs.LG

Junmin Cai , Han Sun , Ningzhong Liu This is my paper · ORCID
classification cs.CVcs.LG
keywords camouflagedobjectdetectionboundaryawarenessreusecross-scalefusionedgeguidancefeatureenhancementPVTv2backboneCODbenchmark
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper argues that boundary-guided camouflaged-object detection fails when edge priors are produced only once, early in the network, because those early priors are noisy and mislead the segmentation. To fix this, B2Net runs a boundary aware module twice: once on enhanced multi-scale features to get an initial boundary, and again after a cross-scale boundary fusion cascade has blended boundary cues into object features, producing a refined boundary that is then concatenated with the fused features. The network is completed by a residual feature enhanced module that enriches each backbone level before the boundary modules see it. The paper reports that this two-pass boundary design outperforms 15 published methods on all four standard metrics on COD10K-Test, Camo-Test, and NC4K-Test, with a three-dataset average $S_\alpha$ gain of 0.93% over the second-best method.

What carries the argument

The load-bearing device is the Boundary Aware Module (BAM), a compact block that adds the two low-level features, multiplies the sum by a high-level semantic feature, and then applies convolutions, a skip connection, max pooling, and spatial attention to keep only object-related edges. The first BAM produces a preliminary edge map; the second BAM, run on the output of the Cross-scale Boundary Fusion Module (CBFM), produces the refined edge map that is concatenated with the fused object features. CBFM carries the fusion: for each shallow level it multiplies the feature by the current edge feature with a learnable weight $\alpha$, concatenates it with the upsampled output of the previous fusion level, and refines the result, so boundary information is propagated downward across scales. A Residual Feature Enhanced Module (RFEM), inspired by Inception and Res2Net, widens each backbone feature with four residual branches before the boundary modules.

What would settle it

Independently re-run the described training recipe (CAMO plus COD10K training subsets, 352x352 inputs, Adam with initial learning rate 8e-5, 100 epochs) and recompute the four metrics on COD10K-Test, Camo-Test, and NC4K-Test; the central claim fails if the reported $S_\alpha$ values of 0.862, 0.866, and 0.882 are not reproduced within a small tolerance.

Watch

Extended reading notes

Core claim

The central claim is that reusing the same boundary-aware module after cross-scale fusion, rather than generating an edge map once at the start, yields both a more accurate boundary and a better camouflaged-object segmentation. In B2Net, the first BAM takes low-level spatial features and high-level semantic features to predict an initial edge; CBFM modules then inject that edge into the object features in a top-down cascade; and a second BAM, reading the fused output, predicts a refined edge that is concatenated with the cascade's features for the final prediction. The paper reports this design reaches $S_\alpha = 0.862$, $0.866$, $0.882$ and $M = 0.023$, $0.048$, $0.033$ on COD10K-Test, Camo-Test, and NC4K-Test respectively, and the ablations attribute the gains to the BAM, the CBFM, and especially the second boundary pass.

Load-bearing premise

The comparison with 15 prior methods is only as strong as the shared protocol behind the baseline numbers in Table I, since the paper provides no code, no error bars, and no significance tests to confirm that the reported one-to-two-point margins would replicate.

Editorial extensions

If this is right

  • The paper's central design principle is that boundary-guided COD should regenerate edge predictions after fusion instead of relying on an early edge prior.
  • Reported gains over the second-best method (average $S_\alpha$ +0.93%, $E_\phi$ +0.68%, $F_\beta^\omega$ +2.5%) indicate that a second boundary pass gives measurable improvements on standard benchmarks.
  • Table III shows the boundary-reuse strategy is transferable: adding a second BAM to BGNet and BSANet improves their scores on the three test sets.
  • The multi-loss supervision (weighted IoU, weighted BCE, and Dice on edges at three side outputs) trains the whole two-boundary network in 100 epochs, so the added complexity is modest.

Reading between the lines

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

  • If the Table I protocol is reproducible, the most direct next stress test is tiny-object scenes, which the paper itself identifies as the main failure mode.
  • The two-pass boundary idea could be tried as a drop-in modification on other COD models that use a single early edge prior; the paper only demonstrates it on BGNet and BSANet.
  • Because the second BAM sees already-fused features, the benefit should be largest when the first edge prior is wrong; a synthetic benchmark with corrupted or degraded boundary supervision could test this prediction.
  • The paper's 352x352 input resolution leaves open whether the boundary-reuse gains persist at higher resolutions where fine edge detail is better preserved.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 6 minor

Summary. The paper proposes B2Net, a camouflaged object detection network built on PVTv2 and composed of a Residual Feature Enhanced Module (RFEM), a Boundary Aware Module (BAM) applied twice, and a Cross-scale Boundary Fusion Module (CBFM). The method is trained on 4,040 images from the CAMO and COD10K training sets and evaluated on COD10K-Test, Camo-Test, and NC4K-Test under four metrics (S_alpha, E_phi, F_beta^w, M). The central claim, stated in Section IV.C and Table I, is that B2Net outperforms 15 published methods on all three datasets under all four metrics. An ablation study (Table II) reports monotone gains from adding each module, and Table III shows that the boundary-reuse strategy improves two existing methods. The paper also includes visual comparisons and failure cases.

Significance. If the reported numbers are reproducible, B2Net makes a modest but legitimate contribution to boundary-guided camouflaged object detection. The idea of reusing a boundary-aware module after feature fusion to refine edge semantics is simple and potentially useful, and the ablation results are internally consistent: each added module improves performance across all three datasets, and the strategy transfers to other backbones or methods (Table III). The paper also provides a failure-case analysis, which is a strength. However, the central quantitative claim is not yet independently verifiable from the manuscript alone: no code, weights, random seeds, or statistical significance tests are provided, and the reported margins over the second-best method FSPNet are as small as 0.4 percentage points in S_alpha on NC4K-Test. These caveats do not imply the method is wrong, but they do mean that the current evidence is not sufficient to fully establish the claimed universal superiority.

major comments (4)
  1. [IV.B] The paper states in Section IV.B that experiments were conducted on four benchmark datasets (CAMO, CHAMELEON, COD10K, NC4K), but Table I reports results only on COD10K-Test, Camo-Test, and NC4K-Test. CHAMELEON is listed but no results are given, and the abstract claims only three datasets. This inconsistency leaves a stated benchmark unreported and makes the claimed 'three popular COD datasets' unclear. Please either report CHAMELEON results or correct the text to three datasets.
  2. [IV.C, Table I] The central claim that B2Net outperforms all 15 methods depends on the comparability of the baseline numbers and on run-to-run variability. The paper does not provide code, weights, random seeds, error bars, or significance tests, and it does not state whether the baseline numbers are copied from the original papers or re-evaluated under the same training and evaluation protocol as B2Net. Since the margins over FSPNet are as small as 0.4 percentage points in S_alpha on NC4K-Test, the ranking could change with seed variation or protocol mismatch. Please supply at least three-seed variance estimates for the main results, and clarify the exact source and protocol of the baseline numbers.
  3. [IV.D] In the 'Effectiveness of CBFM' paragraph, the comparison between M2 and M5 is confounded because M5 differs from M2 by the addition of both CBFM and a second BAM module. The reported improvement in F_beta^w (5.63% on COD10K) cannot therefore be attributed to CBFM alone. Use M1 versus M3 (CBFM alone) or M4 versus M5 (second BAM alone) to isolate each contribution, and rephrase the corresponding claim.
  4. [III.E, Eq. (6)] The loss function in Eq. (6) includes Dice losses on the edge predictions e_i, but the paper does not specify how the edge ground truths G_e are generated during training. Whether they come from dataset-provided edge annotations, from morphological operations on the object masks, or from some other procedure directly affects the BAM supervision and hence the reported boundary quality. Please describe the edge-ground-truth generation process precisely.
minor comments (6)
  1. [Abstract and IV.B] The abstract says 'three challenging benchmark datasets' while Section IV.B says 'four publicly available camouflage object detection benchmark datasets' and lists CHAMELEON. Align the dataset count and the list of used benchmarks.
  2. [IV.B] The text says 'we use five evaluation metrics' but then lists only four (S_alpha, E_phi, F_beta^w, M). Either add the missing metric or correct the count to four.
  3. [III.E] Equation (6) is not typeset clearly in the submitted PDF and appears as an incomplete expression. Ensure the final version has a complete, properly formatted equation with all terms.
  4. [Table I] Some entries have inconsistent decimal precision, e.g., ZoomNet on Camo-Test is listed as S_alpha = 0.82 while other values use three decimals. Use uniform formatting for all table entries.
  5. [References] Reference [3] is cited for both PraNet and ZoomNet, but these are different works (PraNet: Fan et al., MICCAI 2020; ZoomNet: Pang et al., CVPR 2022, reference [13] in the text). Please correct the reference list and citations to avoid ambiguity.
  6. [Fig. 5 caption] The caption mentions 'FAPNet' among the compared methods, but FAPNet does not appear in the method list in Section IV.C or in Table I. Verify the caption and ensure the figure labels match the described comparisons.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: benchmark comparisons, supervised losses, and ablations are external to the claimed result; the main caveats are reproducibility and statistical-support issues, not definitional reductions.

full rationale

The paper's derivation chain is self-contained with respect to the concerns that define circularity. The central result is an empirical comparison against 15 published COD methods on fixed public test splits (COD10K-Test, Camo-Test, NC4K-Test) using standard metrics. The reported B2Net numbers are produced by supervised training with Eq. (6) losses on the CAMO+COD10K training set; the test images and ground truths are external to the model and are not used to fit any parameter. The modules RFEM, BAM, and CBFM are described by explicit architectural equations (1)-(5), and the losses in Eq. (6) do not encode the benchmark outcome except through standard supervised objectives. The ablation study compares variants M1-M5 and reports independent measurements; the M5-vs-M2 comparison is confounded because it changes both CBFM and the second BAM, but that is an experimental confound, not a definitional equivalence. The claim that the reused BAM yields better boundaries is supported by Fig. 7 and the M4-vs-M5 numbers, not by construction. There are no load-bearing self-citations: the authors' prior work is not invoked, and the cited external works (PVTv2, Res2Net, CBAM, datasets, metrics) provide independent components or benchmarks. The main caveats, namely no code, no seeds, no significance tests, CHAMELEON listed but unreported, and thin margins over FSPNet, bear on reproducibility and statistical support, not on circularity. Therefore no circular step is present.

Assumptions & free parameters 2 free parameters · 4 assumptions · 0 invented entities

The central claim is empirical: the reported benchmark scores rest on standard deep learning assumptions (pretrained backbone, reliable labels, comparable baseline numbers) and on the chosen supervision losses. The only named tunable scalar in the architecture is the CBFM fusion weight alpha; the remaining capacity lives in ordinary trained convolution weights, which are not enumerated. No new physical entities are introduced.

free parameters (2)
  • Learnable fusion weight alpha in CBFM = learned during training, not reported
    Eq. 4 scales the element-wise product of edge feature and shallow feature; its learned value affects the fused features and final predictions.
  • Training hyperparameters (LR 8e-5, weight decay 0.1, batch size 16, epochs 100, input size 352x352) = as stated in Section IV.A
    Chosen by hand and central to reproducing the reported results, though they are standard for the field and not fitted to the benchmarks in a derivation sense.
assumptions (4)
  • domain assumption PVTv2 backbone pretrained on ImageNet provides useful multi-scale feature representations for COD.
    The whole architecture consumes backbone features at four scales (Section III.A); no experiment tests an untrained backbone, so reported performance is conditional on these pretrained features.
  • domain assumption Ground-truth camouflaged object masks and boundary maps in the benchmark datasets are correct and consistent.
    Training and all metrics use these labels; the paper does not audit label quality or ambiguity.
  • domain assumption Metric values for the 15 comparison methods are directly comparable to B2Net's numbers.
    Table I reuses published scores; the paper provides no reimplementation, shared evaluation code, or variance information, so comparability of protocols is assumed.
  • domain assumption Dice loss on predicted edge maps is an appropriate supervision signal for boundary quality.
    The loss in Eq. 6 combines weighted BCE/IOU for masks and Dice for edges; the paper does not compare alternative boundary losses.

how reviews work

0 comments
Cite this review

Pith. "Pith review of B2Net: Camouflaged Object Detection via Boundary Aware and Boundary Fusion." pith.science (2026). https://pith.science/paper/QKXQN7IL

@misc{pith2026250100426,
  author       = {Pith},
  title        = {Pith review of: B2Net: Camouflaged Object Detection via Boundary Aware and Boundary Fusion},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/QKXQN7IL}},
  note         = {Machine review of arXiv:2501.00426}
}
read the original abstract

Camouflaged object detection (COD) aims to identify objects in images that are well hidden in the environment due to their high similarity to the background in terms of texture and color. However, existing most boundary-guided camouflage object detection algorithms tend to generate object boundaries early in the network, and inaccurate edge priors often introduce noises in object detection. Address on this issue, we propose a novel network named B2Net aiming to enhance the accuracy of obtained boundaries by reusing boundary-aware modules at different stages of the network. Specifically, we present a Residual Feature Enhanced Module (RFEM) with the goal of integrating more discriminative feature representations to enhance detection accuracy and reliability. After that, the Boundary Aware Module (BAM) is introduced to explore edge cues twice by integrating spatial information from low-level features and semantic information from high-level features. Finally, we design the Cross-scale Boundary Fusion Module(CBFM) that integrate information across different scales in a top-down manner, merging boundary features with object features to obtain a comprehensive feature representation incorporating boundary information. Extensive experimental results on three challenging benchmark datasets demonstrate that our proposed method B2Net outperforms 15 state-of-art methods under widely used evaluation metrics. Code will be made publicly available.

Figures

Figures reproduced from arXiv: 2501.00426 by the authors.

Figure 1
Figure 1. The overall architecture of the proposed B2Net, which consists of three key components, i.e., Residual Feature Enhanced Module (RFEM), Boundary Aware Module(BAM) and Cross-scale Boundary Fusion Module(CBFM) [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. The detailed architecture of the proposed Residual Feature Enhanced Module(RFEM). C. Boundary Aware Module Boundary information is of great significance for the segmentation and localization of camouflaged objects. Camouflage images have high resolution in low-level feature regions and can extract boundary information, but they are inevitably interfered with by the boundary information of non-camouflaged objects in … view at source ↗
Figure 1
Figure 1. shows the overview of the proposed B 2Net, which consists of three kinds of key components including Residual Feature Enhanced Module, Boundary Aware Module, and Cross-scale Boundary Fusion Module. Specifically, given an input image I, we first adopt the PVTv2 [11] as backbone to extract features at four levels, which can be denoted as F {f ,i 1,2,3,4}  i . Then, we feed F into Residual Feature Enhanced Module to … view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

32 extracted references · 31 canonical work pages

  1. [1]

    Concealed object detection

    Fan, Deng -Ping, Ge -Peng Ji, Ming -Ming Cheng, and Ling Shao. "Concealed object detection." IEEE transactions on pattern analysis and machine intelligence 44, no. 10 (2021): 6024-6042

    work page 2021

  2. [2]

    Animal camouflage: current issues and new perspectives

    Stevens, Martin, and Sami Merilaita. "Animal camouflage: current issues and new perspectives." Philosophical Transactions of the Royal Society B: Biological Sciences 364, no. 1516 (2009): 423-427

    work page 2009

  3. [3]

    Pranet: Parallel reverse attention network for polyp segmentation

    Fan, Deng-Ping, Ge -Peng Ji, Tao Zhou, Geng Chen, Huazhu Fu, Jianbing Shen, and Ling Shao. "Pranet: Parallel reverse attention network for polyp segmentation." In International conference on medical image computing and compute r-assisted intervention, pp. 263 -273. Cham: Springer International Publishing, 2020

    work page 2020

  4. [4]

    Marine animal segmentation

    Li, Lin, Bo Dong, Eric Rigall, Tao Zhou, Junyu Dong, and Geng Chen. "Marine animal segmentation." IEEE Transactions on Circuits and Systems for Video Technology 32, no. 4 (2021): 2303-2314

    work page 2021

  5. [5]

    Camouflage images

    Chu, Hung-Kuo, Wei-Hsin Hsu, Niloy J. Mitra, Daniel Cohen -Or, Tien-Tsin Wong, and Tong -Yee Lee. "Camouflage images." ACM Trans. Graph. 29, no. 4 (2010): 51-1

    work page 2010

  6. [6]

    Camouflaged object detection

    Fan, Deng -Ping, Ge -Peng Ji, Guolei Sun, Ming -Ming Cheng, Jianbing Shen, and Li ng Shao. "Camouflaged object detection." In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pp. 2777-2787. 2020

    work page 2020

  7. [7]

    Camouflaged object detection via context-aware cross-level fusion

    Chen, Geng, Si -Jie Liu, Yu -Jia Sun, Ge -Peng Ji, Ya -Feng Wu, and Tao Zhou. "Camouflaged object detection via context-aware cross-level fusion." IEEE Transactions on Circuits and Systems for Video Technology 32, no. 10 (2022): 6981-6993

    work page 2022

  8. [8]

    Deep texton-coherence network for camouflaged object detection

    Zhai, Wei, Yang Cao, HaiYong Xie, and Zheng -Jun Zha. "Deep texton-coherence network for camouflaged object detection." IEEE Tran sactions on Multimedia 25 (2022): 5155-5165

    work page 2022

Show all 32 references
  1. [9]

    Boundary-guided camouflaged object detection

    Sun, Yujia, Shuo Wang, Chenglizhao Chen, and Tian -Zhu Xiang. "Boundary-guided camouflaged object detection." arXiv preprint arXiv:2207.00794 (2022)

    2022 arXiv

  2. [10]

    I can find you! boundary-guided separated attention network for camouflaged object detection

    Zhu, Hongwei, Peng Li, Haoran Xie, Xuefeng Yan, Dong Liang, Dapeng Chen, Mingqiang Wei, and Jing Qin. "I can find you! boundary-guided separated attention network for camouflaged object detection." In Proceedings of the AAAI conference on artificial intelligence, vol. 36, no. ...

    2022

  3. [11]

    Pvt v2: Improved baselines with pyramid vision transformer

    Wang, Wenhai, Enz e Xie, Xiang Li, Deng -Ping Fan, Kaitao Song, Ding Liang, Tong Lu, Ping Luo, and Ling Shao. "Pvt v2: Improved baselines with pyramid vision transformer." Computational Visual Media 8, no. 3 (2022): 415-424

    2022

  4. [12]

    Preynet: Preying on camouflaged objects

    Zhang, Miao, Shuang Xu, Yongri Piao, Dongxiang Shi, Shusen Lin, and Huchuan Lu. "Preynet: Preying on camouflaged objects." In Proceedings of the 30th ACM international conference on multimedia, pp. 5323-5332. 2022

    2022

  5. [13]

    Zoom in and out: A mixed-scale triplet network for camouflaged object detection

    Pang, Youwei, Xiaoqi Zhao, Tian -Zhu Xiang, Lihe Zhang, and Huchuan Lu. "Zoom in and out: A mixed-scale triplet network for camouflaged object detection." In Proceedings of the IEEE/CVF Conference on computer vision and pattern recognition, pp. 2160-2170. 2022

    2022

  6. [14]

    Edge -aware mirror network for camouflaged obj ect detection

    Sun, Dongyue, Shiyao Jiang, and Lin Qi. "Edge -aware mirror network for camouflaged obj ect detection." In 2023 IEEE International Conference on Multimedia and Expo (ICME) , pp. 2465-2470. IEEE, 2023

    2023

  7. [15]

    Res2net: A new multi -scale backbone architecture

    Gao, Shang -Hua, Ming -Ming Cheng, Kai Zhao, Xin -Yu Zhang, Ming-Hsuan Yang, and Philip Torr. "Res2net: A new multi -scale backbone architecture." IEEE transactions on pattern analysis and machine intelligence 43, no. 2 (2019): 652-662

    2019

  8. [16]

    Cbam: Convolutional block attention module

    Woo, Sanghyun, Jongchan Park, Joon -Young Lee, and In So Kweon. "Cbam: Convolutional block attention module." In Proceedings of the European conference on computer visi on (ECCV), pp. 3-19. 2018

    2018

  9. [17]

    Multi-scale interactive network for salient object detection

    Pang, Youwei, Xiaoqi Zhao, Lihe Zhang, and Huchuan Lu. "Multi-scale interactive network for salient object detection." In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pp. 9413-9422. 2020

    2020

  10. [18]

    Anabranch network for camouflaged object segmentation

    Le, Trung-Nghia, Tam V. Nguyen, Zhongliang Nie, Minh -Triet Tran, and Akihiro Sugimoto. "Anabranch network for camouflaged object segmentation." Computer vision and image understanding 184 (2019): 45-56

    2019

  11. [19]

    Animal camouflage analysis: Chameleon database

    Skurowski, Przemysław, Hassan Abdulameer, J. Błaszczyk, Tomasz Depta, Adam Kornacki, and P. Kozieł. "Animal camouflage analysis: Chameleon database." Unpublished manuscript 2, no. 6 (2018): 7

    2018

  12. [20]

    Simultaneously localize, segment and rank the camouflaged objects

    Lv, Yunqiu, Jing Zhang, Yuchao Dai, Aixuan Li, Bowen Liu, Nick Barnes, and Deng -Ping Fan. "Simultaneously localize, segment and rank the camouflaged objects." In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pp. 11591-11601. 2021

    2021

  13. [21]

    Cognitive vision inspired object segmentation metric and loss function

    Fan, Deng-Ping, Ge-Peng Ji, Xuebin Qin, and Ming -Ming Cheng. "Cognitive vision inspired object segmentation metric and loss function." Scientia Sinica Informationis 6, no. 6 (2021): 5

    2021

  14. [22]

    Structure-measure: A new way to evaluate foreground maps

    Fan, Deng-Ping, Ming-Ming Cheng, Yun Liu, Tao Li, and Ali Borji. "Structure-measure: A new way to evaluate foreground maps ." In Proceedings of the IEEE international conference on computer vision, pp. 4548-4557. 2017

    2017

  15. [23]

    How to evaluate foreground maps?

    Margolin, Ran, Lihi Zelnik -Manor, and Ayellet Tal. "How to evaluate foreground maps?." In Proceedings of the IEEE conference on computer vision and pattern reco gnition, pp. 248-255. 2014

    2014

  16. [24]

    Saliency filters: Contrast based filtering for salient region detection

    Perazzi, Federico, Philipp Krähenbühl, Yael Pritch, and Alexander Hornung. "Saliency filters: Contrast based filtering for salient region detection." In 2012 IEEE conference on computer vision and pattern recognition, pp. 733-740. IEEE, 2012

    2012

  17. [25]

    Context-aware cross -level fusion network for camouflaged object detection

    Sun, Yujia, Geng Chen, Tao Zhou, Yi Zhang, and Nian Liu. "Context-aware cross -level fusion network for camouflaged object detection." arXiv preprint arXiv:2105.12555 (2021)

    2021 arXiv

  18. [26]

    Camouflaged object segmentation with distraction mining

    Mei, Haiyang, Ge-Peng Ji, Ziqi Wei, Xin Yang, Xiaopeng Wei, and Deng-Ping Fan. "Camouflaged object segmentation with distraction mining." In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pp. 8772 -8781. 2021

    2021

  19. [27]

    Camouflaged Object Detection with Feature Grafting and Distractor Aware

    Song, Yuxuan, Xinyue Li, and Lin Qi. "Camouflaged Object Detection with Feature Grafting and Distractor Aware." In 2023 IEEE International Conference on Multimedia and Expo (ICME) , pp. 2459-2464. IEEE, 2023

    2023

  20. [28]

    Camouflaged object detection with f eature decomposition and edge reconstruction

    He, Chunming, Kai Li, Yachao Zhang, Longxiang Tang, Yulun Zhang, Zhenhua Guo, and Xiu Li. "Camouflaged object detection with f eature decomposition and edge reconstruction." In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pp. 22046-22055. 2023

    2023

  21. [29]

    Depth -aided camouf laged object detection

    Wang, Qingwei, Jinyu Yang, Xiaosheng Yu, Fangyi Wang, Peng Chen, and Feng Zheng. "Depth -aided camouf laged object detection." In Proceedings of the 31st ACM International Conference on Multimedia, pp. 3297-3306. 2023

    2023

  22. [30]

    Feature shrinkage pyramid for camouflaged object detection with transformers

    Huang, Zhou, Hang Dai, Tian -Zhu Xiang, Shuo Wang, Huai -Xin Chen, Jie Qin, and Huan Xiong. "Feature shrinkage pyramid for camouflaged object detection with transformers." In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pp. 5557-5566. 2023

    2023

  23. [31]

    Camouflaged object detection via cross -level refinement and interaction network

    Ge, Yanliang, Junchao Ren, Qiao Zhang, Min He, Hongbo Bi, and Cong Zhang. "Camouflaged object detection via cross -level refinement and interaction network." Image and Vision Computing 144 (2024): 104973

    2024

  24. [32]

    Local to global purification strategy to realize collaborative camouflaged object detection

    Tong, Jinghui, Yaqiu Bi, Cong Zhang, Hongbo Bi, and Ye Yuan. "Local to global purification strategy to realize collaborative camouflaged object detection." Computer Visio n and Image Understanding 241 (2024): 103932

    2024

Pith tools

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