Pith. sign in

REVIEW 4 major objections 6 minor 67 references

VCP-DCN: Beyond Visual Concealed Property via Depth Collaborative Network for Camouflaged Object Detection

T0 review · 4 major / 6 minor · reviewed 2026-08-01 · deepseek-v4-flash

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

Pith's one-line read The paper claims that a depth collaborative network, which explicitly learns modality-consistency and modality-specific prototypes via contrastive learning, achieves state-of-the-art camouflaged object detection by preventing RGB and depth

desk verdict A coherent RGB-D COD architecture with a solid ablation story, but the headline SOTA claim rests on a COD10K Fwβ number that looks like a typo and needs verification before publication. read the letter →

arxiv 2607.27843 v1 pith:P5H5FJJP submitted 2026-07-30 cs.CV

Songsong Duan , Xi Yang , Nannan Wang This is my paper
classification cs.CV
keywords camouflagedobjectdetectionRGB-Dfusionmulti-modalitylearningprototypecontrastivedepthestimationvisualstatespacemodelimagesegmentation
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 addresses a core failure it identifies in existing RGB-D camouflaged object detection: when RGB and depth features are fused under a single mask loss, the modalities converge into homogeneous representations and lose their complementary strengths. The authors propose VCP-DCN, a depth collaborative network that progressively aligns, interacts, and fuses RGB and depth features while explicitly preserving what is shared and what is unique to each modality. They introduce prototype contrastive learning to the COD task, separating RGB and depth features into modality-consistency and modality-specific prototypes. If the claims hold, VCP-DCN becomes the new state of the art on three standard camouflaged-object benchmarks and also transfers to RGB-D salient object detection.

What carries the argument

The central object is a set of learned prototype tokens—foreground and background, further separated into modality-consistency and modality-specific variants—derived from RGB and depth features via the SPE module. These prototypes are trained with four contrastive losses (modality-consistency, modality-specific, and two foreground-background separation losses). The MDA module uses the consistency prototypes to generate foreground and background masks, then applies masked linear attention to exchange cross-modal cues. The DAI module computes cosine similarities between specific and consistency prototypes to adaptively weight RGB and depth contributions during fusion. The entire mechanism is d

What would settle it

Train the same pipeline on the same benchmarks with real depth maps (e.g., from a depth sensor) or with corrupted pseudo-depth and compare; if performance drops to RGB-only levels, the claimed depth benefit is an artifact of the specific estimator.

Watch

Extended reading notes

Core claim

VCP-DCN outperforms previous state-of-the-art COD methods on CAMO, COD10K, and NC4K, and it does so while using relatively few parameters and FLOPs. The central innovation is a progressive three-stage pipeline: the Separable Prototype Embedding (SPE) module uses prototype contrastive losses to separate modality-consistency and modality-specific prototypes for foreground and background; the Multi-modality Dual Attention (MDA) module exchanges cross-modal prototype-derived masks to refine features; and the Depth Adaptive Injection (DAI) module weighs RGB and depth contributions based on similarity between specific and consistency prototypes. The paper also shows the same network transfers to R

Load-bearing premise

The results depend on the pseudo-depth maps generated by a monocular depth estimator being informative enough about geometry for camouflaged scenes; if those maps are noisy or uninformative, the depth branch contributes no signal.

Editorial extensions

If this is right

  • RGB-D camouflaged object detection surpasses RGB-only methods on CAMO, COD10K, and NC4K, establishing depth as a reliable complementary cue for concealed objects.
  • Prototype contrastive learning prevents multimodal feature homogenization, directly addressing a bias the paper identifies in earlier RGB-D fusion approaches.
  • The same depth collaborative network transfers to RGB-D salient object detection, suggesting a general mechanism for RGB-D segmentation beyond camouflage.
  • With 60.2M parameters and 46.4G FLOPs, the architecture is efficient enough for real-time applications while maintaining high accuracy.

Reading between the lines

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

  • Editorial inference: the paper does not test robustness to depth errors; a natural extension would be to measure performance as pseudo-depth noise increases or with real depth sensors, to see if the depth benefit persists.
  • Editorial inference: the separable prototype space could serve as an interpretability tool, revealing what depth contributes beyond appearance for each camouflaged scene.
  • Editorial inference: the prototype contrastive mechanism may transfer to other modality pairs, such as RGB-thermal or RGB-multispectral, for concealed object detection in different domains.
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 / 6 minor

Summary. The paper proposes VCP-DCN, an RGB-D camouflaged object detection (COD) method that introduces a three-stage depth collaborative network: a Separable Prototype Embedding (SPE) module for aligning RGB and depth features through modality-consistency and modality-specific prototype contrastive learning, a Multi-modality Dual Attention (MDA) module for cross-modal feature interaction, and a Depth Adaptive Injection (DAI) module for adaptive fusion. The method is evaluated on CAMO, COD10K, and NC4K with pseudo-depth from Depth Anything V2, and the authors claim consistent state-of-the-art performance over existing RGB and RGB-D COD methods, with additional experiments on RGB-D salient object detection (SOD). The architecture is coherent and the ablations indicate that each component contributes, but the central quantitative claim is undermined by an unexplained and internally inconsistent set of Fwβ scores in Table 1.

Significance. If the reported results are reliable, VCP-DCN would represent a genuine advance in RGB-D COD, demonstrating the value of explicitly separating modality-consistency and modality-specific prototypes in a depth collaborative framework. The extension to RGB-D SOD also suggests that the modules are transferable. The design is well motivated and the ablation study supports the utility of the three proposed modules. However, the paper's headline conclusion hinges on Table 1, where a suspicious COD10K Fwβ value (0.930/0.932) is far outside the range of all comparable methods and is not corroborated by other metrics. The lack of reproducibility details (no commit hash, no evaluation config, no multiple seeds) further weakens confidence. The central claim is therefore currently unverified, and the manuscript requires substantial revision to either justify or correct the anomalous numbers.

major comments (4)
  1. [Table 1, Sec. 4.2] In the 'RGB-D-based COD Methods via Depth Anything V2' block of Table 1, VCP-DCN reports COD10K Fwβ = 0.930 (VMamba) and 0.932 (Swin). This is a ~0.09 gain over the same-backbone Samba row (0.842) and ~0.13 over DaCOD (0.795), while the accompanying improvements in Sm, Eϕ, and M are much smaller (e.g., Sm: 0.881 vs 0.870; M: 0.017 vs 0.019). The Fwβ value 0.930 also exceeds the reported CAMO/NC4K Fwβ (0.871) and is higher than the COD10K Sm (0.881), which is atypical for this benchmark. No ablation in Table 3 produces a gain larger than 0.03 Fwβ. These numbers are load-bearing for the 'consistently outperforms' claim, and the paper provides no error bars, multiple seeds, or runnable evaluation configuration to verify them. Please check for typographical errors or column misalignment and provide corrected or verified values.
  2. [Sec. 3.2, Eqs. (3)-(5)] The prototype contrastive losses L_MCL, L_MSL, L_RCL, and L_DCL compute logarithms of sums or differences of cosine similarities without clamping, a temperature, or a softmax formulation. Since cosine similarity lies in [-1,1], the argument of the logarithm (e.g., Cos(F_P_rc, F_P_dc) + Cos(B_P_r, B_P_d) in Eq. (3), or 1 - Cos(F_P_rs, F_P_ds) in Eq. (4)) can be zero or negative at initialization, making the loss undefined or infinite. The paper does not describe any numerical stabilization (e.g., epsilon or temperature). This is central to the SPE module, and the authors should clarify the exact implementation or provide a stable variant with a reference to the standard contrastive-learning form.
  3. [Sec. 4.1 and Code Release] The code link is given in the abstract, but no commit hash, evaluation scripts, or training/validation configuration are provided. The numerical claims, especially the anomalous COD10K Fwβ values, cannot be independently reproduced without these artifacts. The authors should release the exact evaluation harness, the trained models or multiple random seeds, and the pseudo-depth generation pipeline. Without this, the central comparison table is not verifiable.
  4. [Sec. 4.1 and Table 1 (depth-source comparison)] All experiments rely on pseudo-depth from Depth Anything V2; the only depth-source comparison is DPT vs. Depth Anything V2 in Table 1. There is no sensitivity analysis to depth noise or to real sensor depth. Since the method's core contribution is a 'depth collaborative network,' it is important to show that the gains are not an artifact of the specific pseudo-depth estimator. Adding an experiment with noisy or alternative depth input (e.g., corrupting Depth Anything V2 outputs or using a lower-quality depth estimator) would directly address this robustness concern.
minor comments (6)
  1. [References] Several references are duplicated: [7] and [8] are the same paper; [10] and [11]; [23] and [24]; [50] and [51]. Please consolidate. This is particularly distracting for the authors' own prior work ([7]/[8]).
  2. [Sec. 4.3] Ablation studies are reported only on CAMO. The title claims 'consistently outperforms' across three datasets; reporting ablations on at least COD10K or NC4K would strengthen that claim.
  3. [Table 1] The RGB-based group includes Samba, while the RGB-D-based DPT group also includes Samba. The grouping is confusing; please clarify whether Samba is used with RGB-only in one group and with depth in the other, or rename the rows to avoid ambiguity.
  4. [Sec. 3.4, Eq. (9)] Eq. (9) computes similarity scores only for the foreground modality-specific prototypes (F_P_rs, F_P_ds) against the consistency prototype. For symmetry with Eq. (10), where both foreground and background specific features are used in fusion, the paper should also define background-specific similarity scores or explain why only foreground prototypes are used in the decision-making mechanism.
  5. [Fig. 7(a)] The caption says 'we visualize the semantic maps SM_r^3 and SM_r^3'; the second should presumably be SM_d^3 (depth branch).
  6. [Sec. 5 and Sec 4.3] Minor language issues: 'we presents VCP-DCN' in Sec. 5 and 'boots' in Sec. 4.3 ('F-MLA and B-MLA boots A1 by 9.8%') should be 'present' and 'boost', respectively.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the network is supervised on held-out benchmarks and no equation reduces to its inputs by construction.

full rationale

VCP-DCN is a standard supervised RGB-D COD architecture. The three proposed modules (SPE, MDA, DAI) operate on encoder features and are optimized by the total loss in Eq. (11) against ground-truth masks on the training split; no module output is, by construction, equal to a fitted parameter renamed as a prediction. The prototype contrastive losses in Eqs. (3)-(5) regularize learned prototypes but do not encode the test-set metrics. The only pseudo-depth input, Depth Anything V2, is an external pretrained model cited as [56]; using it is a data-source choice, not a circular derivation. Tables 1 and 2 evaluate held-out test sets, so the SOTA claim is empirical rather than derived from the inputs. The duplicated self-citations (e.g., [7]/[8], [9]/[10], [13]/[14]) are bibliography errors; the triple-decoder design is attributed to [8] and [66], but this design choice is not the paper's central claim and the external citation [66] provides independent support. A suspicious COD10K Fwβ value (0.930/0.932) is an internal-consistency or reproducibility concern, not evidence of circularity.

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

The central claim rests on standard deep-learning assumptions plus the premise that generated depth maps are informative and that prototype contrastive separation captures modality differences. Two loss weights, λ and μ, are hand-set. No new physical or ontological entities are proposed; prototype tokens are internal learned representations.

free parameters (2)
  • λ (mask loss weight) = 0.3
    Chosen by hand in Section 3.5/4.1 to balance mask and alignment losses; no sensitivity analysis reported.
  • μ (alignment loss weight) = 0.1
    Chosen by hand in Section 3.5/4.1 to weight the prototype contrastive alignment loss; no sensitivity analysis reported.
assumptions (4)
  • domain assumption Depth maps produced by Depth Anything V2 provide reliable geometry cues for camouflaged objects
    Adopted in Section 4.1 to generate pseudo-depth; if depth is uninformative, the proposed depth branch cannot help.
  • ad hoc to paper Modality-consistency and modality-specific prototypes are separable by cosine-distance contrastive losses
    SPE training relies on this assumption in Section 3.2 (Eqs. 3-5); the paper provides t-SNE and heatmap evidence but no formal guarantee.
  • domain assumption Standard supervised-learning assumption that train/test images are i.i.d. from the same benchmark distributions and that the evaluation splits are comparable to prior works
    Section 4.1; the fairness of the SOTA comparison depends on identical split and protocol.
  • domain assumption Backbone VMamba-S/Swin-S pretrained on ImageNet provides features that transfer to COD
    Used throughout the encoder; standard transfer-learning assumption not verified for this specific task.

how reviews work

0 comments
Cite this review

Pith. "Pith review of VCP-DCN: Beyond Visual Concealed Property via Depth Collaborative Network for Camouflaged Object Detection." pith.science (2026). https://pith.science/paper/P5H5FJJP

@misc{pith2026260727843,
  author       = {Pith},
  title        = {Pith review of: VCP-DCN: Beyond Visual Concealed Property via Depth Collaborative Network for Camouflaged Object Detection},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/P5H5FJJP}},
  note         = {Machine review of arXiv:2607.27843}
}
read the original abstract

Camouflaged Object Detection (COD) aims to identify and segment camouflaged objects in complex environments, which are often concealed because their color and texture are similar to the background. Several existing COD methods introduce depth maps to boost detection performance via learning complementary RGB-D features, ignoring modality-specific characteristics of concealed objects in the depth domain. To address this issue, we propose a depth collaborative network, called VCP-DCN, to mine distinguishable multi-modality features beyond visual concealed prototype in depth domain. Specifically, VCP-DCN progressively performs multi-modality alignment, interaction, and fusion for the COD task. In the \textbf{alignment} stage, we propose a Separable Prototype Embedding (SPE) module to learn modality-consistency and modality-specific RGB/depth prototype tokens through prototype contrastive learning. Furthermore, we develop a Multi-modality Dual Attention (MDA) module to enhance the cross-modal feature representation through local response maps between modality-consistency RGB/depth prototype tokens and visual tokens on the \textbf{interaction} stage. Finally, we design a Depth Adaptive Injection (DAI) module to adaptively measure contribution of RGB/depth features with a decision-making mechanism, which calculates similarity distance between RGB/depth modality-specific prototype tokens and modality-consistency ones on the \textbf{fusion} stage. Extensive experiments demonstrate the effectiveness of our VCP-DCN on three authoritative datasets.

Figures

Figures reproduced from arXiv: 2607.27843 by the authors.

Figure 1
Figure 1. Comparison with RGB and Depth images via histogram and attention maps. Consistent RGB-D feature RGB-specific feature Depth-specific feature Mask Loss RGB+D Fusion RGB Encoder Depth Encoder (a) Existing Methods RGB features Depth features RGB+D Fusion Prototype Space Contrastive Loss RGB features Depth features RGB Encoder Depth Encoder Refine Mask Loss (b) Multi-modality Distribution (c) Our VCP-DCN [PITH_FULL_IMAG… view at source ↗
Figure 2
Figure 2. Comparison with existing RGB-D COD and our VCP-DCN. (a) is the existing RGB-D COD methods, which learns the RGB-D complementary features under the supervision of mask loss; (b) is the illustration of multi-modality feature distribution; (c) is our VCP-DCN, which learns multi-modality specific features to refine the RGB-D features via separable prototype learning. attributes to blend into their surrounding environmen… view at source ↗
Figure 3
Figure 3. Illustration of the proposed VCP-DCN, which consists of Separable Prototype Embedding (SPE), Multi-modality Dual Attention (MDA), Depth Adaptive Injection (DAI) modules. In the alignment stage, a prototype contrastive loss (Lalign) is proposed to lean modality-consistency and modality-specific prototypes, which contains LMCL, LMSL, and LRCL. decoders to obtains prediction masks Mr, Md, and Mf . To achieve an effec￾t… view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: Illustration of MDA module of RGB feature and DAI module. Depth feature shares a similar MDA structure. Under the constraints of LMCL and LMSL, we learn modality-consistency and modality-specific representation by optimizing the similarity distance be￾tween multi-modal…
Figure 5
Figure 5. Figure 5: Visual comparisons between our VCP-DCN and several previous SOTA meth￾ods, including Samba [14], CamoDiffusion [41], FSEL [42], and CamoFormer [57] on challenging scenes. Method NJU2K NLPR DUT F w β ↑ M ↓ F w β ↑ M ↓ F w β ↑ M ↓ AFB [44] 0.910 0.028 0.902 0.021 0.919 0…
Figure 6
Figure 6. Figure 6: Efficiency comparisons between our VCP-DCN and several SOTA methods on CAMO dataset. Item SPE MDA DAI CAMO Sm Eϕ Fw β ↑ M ↓ C1 0.856 0.907 0.793 0.052 C2 ✓ 0.871 0.925 0.822 0.045 C3 ✓ ✓ 0.883 0.931 0.854 0.039 C4 ✓ ✓ 0.881 0.934 0.841 0.041 C5 ✓ ✓ ✓ 0.892 0.946 0.871 …
Figure 7
Figure 7. Figure 7: Heatmap visualization of multi-modality semantic maps SM3 r and SM3 d in SPE module. Effectiveness of VCP-DCN. To investigate the role of the depth collabora￾tive network in our \mathcal {VCP\text {-}DCN} , we conduct faithful ablation studies in [PITH_FULL_IMAGE:figu…

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

67 extracted references · 3 linked inside Pith

  1. [8]

    IEEE Transactions on Image Process- ing (2025)

    Duan, S., Yang, X., Wang, N., Gao, X.: Lightweight rgb-d salient object detection from a speed-accuracy tradeoff perspective. IEEE Transactions on Image Process- ing (2025)

    2025

  2. [11]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition

    Fan, D.P., Ji, G.P., Sun, G., Cheng, M.M., Shen, J., Shao, L.: Camouflaged object detection. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. pp. 2777–2787 (2020)

    2020

  3. [13]

    In: Proceedings of the Computer Vision and Pat- tern Recognition Conference

    He, J., Fu, K., Liu, X., Zhao, Q.: Samba: A unified mamba-based framework for general salient object detection. In: Proceedings of the Computer Vision and Pat- tern Recognition Conference. pp. 25314–25324 (2025)

    2025

  4. [14]

    In: Proceedings of the Computer Vision and Pat- tern Recognition Conference

    He, J., Fu, K., Liu, X., Zhao, Q.: Samba: A unified mamba-based framework for general salient object detection. In: Proceedings of the Computer Vision and Pat- tern Recognition Conference. pp. 25314–25324 (June 2025)

    2025

  5. [24]

    In: Proceedings of the IEEE/CVF International Conference on Computer Vision (2025)

    Liu, J., Kong, L., Chen, G.: Improving sam for camouflaged object detection via dual stream adapters. In: Proceedings of the IEEE/CVF International Conference on Computer Vision (2025)

    2025

  6. [46]

    In: Proceedings of the ACM International Conference on Multimedia

    Wang, Q., Yang, J., Yu, X., Wang, F., Chen, P., Zheng, F.: Depth-aided camou- flaged object detection. In: Proceedings of the ACM International Conference on Multimedia. pp. 3297–3306 (2023)

    2023

  7. [47]

    In: Proceedings of the ACM international Conference on Multimedia

    Wang, Q., Yang, J., Yu, X., Wang, F., Chen, P., Zheng, F.: Depth-aided camou- flaged object detection. In: Proceedings of the ACM international Conference on Multimedia. pp. 3297–3306 (2023) VCP-DCN 19

    2023

  8. [51]

    In: Proceedings of the IEEE/CVF International Conference on Computer Vision

    Wu, Z., Paudel, D.P., Fan, D.P., Wang, J., Wang, S., Demonceaux, C., Timofte, R., Van Gool, L.: Source-free depth for object pop-out. In: Proceedings of the IEEE/CVF International Conference on Computer Vision. pp. 1032–1042 (2023)

    2023

  9. [62]

    International Journal of Computer Vision pp

    Zhang, D., Cheng, L., Liu, Y., Wang, X., Han, J.: Mamba capsule routing to- wards part-whole relational camouflaged object detection. International Journal of Computer Vision pp. 1–21 (2025)

    2025

  10. [63]

    International Journal of Computer Vision pp

    Zhang, D., Cheng, L., Liu, Y., Wang, X., Han, J.: Mamba capsule routing to- wards part-whole relational camouflaged object detection. International Journal of Computer Vision pp. 1–21 (2025) 20 Duan et al

    2025

Show all 67 references
  1. [1]

    IEEE Transac- tions on Neural Networks and Learning Systems36(2), 3175–3188 (2024)

    Chen, G., Wang, Q., Dong, B., Ma, R., Liu, N., Fu, H., Xia, Y.: Em-trans: Edge- aware multimodal transformer for rgb-d salient object detection. IEEE Transac- tions on Neural Networks and Learning Systems36(2), 3175–3188 (2024)

    2024

  2. [2]

    IEEE Transactions on Industrial Informatics (2025)

    Cheng, H., Luo, J., Zhang, X.: Multimodal industrial anomaly detection via uni- modal and cross-modal fusion. IEEE Transactions on Industrial Informatics (2025)

    2025

  3. [3]

    In: Proceedings of the ACM International Conference on Multimedia

    Cong, R., Liu, H., Zhang, C., Zhang, W., Zheng, F., Song, R., Kwong, S.: Point- aware interaction and cnn-induced refinement network for rgb-d salient object de- tection. In: Proceedings of the ACM International Conference on Multimedia. pp. 406–416 (2023)

    2023

  4. [4]

    In: Proceedings of the ACM International Conference on Multimedia

    Cong, R., Sun, M., Zhang, S., Zhou, X., Zhang, W., Zhao, Y.: Frequency per- ception network for camouflaged object detection. In: Proceedings of the ACM International Conference on Multimedia. pp. 1179–1189 (2023)

    2023

  5. [5]

    Nature434(7029), 72–74 (2005)

    Cuthill, I.C., Stevens, M., Sheppard, J., Maddocks, T., Párraga, C.A., Troscianko, T.S.: Disruptive coloration and background pattern matching. Nature434(7029), 72–74 (2005)

    2005

  6. [6]

    arXiv preprint arXiv:2108.06932 (2021)

    Dong, B., Wang, W., Fan, D.P., Li, J., Fu, H., Shao, L.: Polyp-pvt: Polyp segmen- tation with pyramid vision transformers. arXiv preprint arXiv:2108.06932 (2021)

    2021 arXiv

  7. [9]

    IEEE Transactions on Pattern Analysis and Machine Intelligence44(10), 6024–6042 (2021)

    Fan, D.P., Ji, G.P., Cheng, M.M., Shao, L.: Concealed object detection. IEEE Transactions on Pattern Analysis and Machine Intelligence44(10), 6024–6042 (2021)

    2021

  8. [12]

    In: Proceed- ings of the ACM International Conference on Multimedia

    Hao, Z., Xiao, Z., Luo, Y., Guo, J., Wang, J., Shen, L., Hu, H.: Primkd: Primary modality guided multimodal fusion for rgb-d semantic segmentation. In: Proceed- ings of the ACM International Conference on Multimedia. pp. 1943–1951 (2024)

    1943

  9. [15]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition

    He, K., Zhang, X., Ren, S., Sun, J.: Deep residual learning for image recognition. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. pp. 770–778 (2016)

    2016

  10. [16]

    Huang, Z., Dai, H., Xiang, T.Z., Wang, S., Chen, H.X., Qin, J., Xiong, H.: Feature shrinkagepyramidforcamouflagedobjectdetectionwithtransformers.In:Proceed- ings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. pp. 5557–5566 (2023) VCP-DCN 17

    2023

  11. [17]

    IEEE Transactions on Image Processing31, 2321–2336 (2022)

    Ji, W., Yan, G., Li, J., Piao, Y., Yao, S., Zhang, M., Cheng, L., Lu, H.: Dmra: Depth-induced multi-scale recurrent attention network for rgb-d saliency detection. IEEE Transactions on Image Processing31, 2321–2336 (2022)

    2022

  12. [18]

    In: IEEE International Conference on Image Process- ing

    Ju, R., Ge, L., Geng, W., Ren, T., Wu, G.: Depth saliency based on anisotropic center-surround difference. In: IEEE International Conference on Image Process- ing. pp. 1115–1119. IEEE (2014)

    2014

  13. [19]

    In: Proceedings of the IEEE/CVF International Conference on Computer Vision

    Kirillov, A., Mintun, E., Ravi, N., Mao, H., Rolland, C., Gustafson, L., Xiao, T., Whitehead, S., Berg, A.C., Lo, W.Y., et al.: Segment anything. In: Proceedings of the IEEE/CVF International Conference on Computer Vision. pp. 4015–4026 (2023)

    2023

  14. [20]

    Computer Vision and Image Understanding 184, 45–56 (2019)

    Le, T.N., Nguyen, T.V., Nie, Z., Tran, M.T., Sugimoto, A.: Anabranch network for camouflaged object segmentation. Computer Vision and Image Understanding 184, 45–56 (2019)

    2019

  15. [21]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition

    Li, Y., Zhu, Z., Zhang, Y., Chen, Y., Yu, Z.: Boost the inference with co-training: A depth-guided mutual learning framework for semi-supervised medical polyp seg- mentation. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. pp. 10394–10403 (2025)

    2025

  16. [22]

    In: Proceedings of the International Conference on Machine Learning

    Lian, S., Zhang, Z., Li, H., Li, W., Yang, L.T., Kwong, S., Cong, R.: Diving into underwater: Segment anything model guided underwater salient instance segmen- tation and a large-scale dataset. In: Proceedings of the International Conference on Machine Learning. pp. 29545–29559 (2024)

    2024

  17. [25]

    IEEE Transactions on Pattern Analysis and Machine Intelligence 46(11), 7300–7316 (2025)

    Liu, N., Luo, Z., Zhang, N., Han, J.: Vst++: Efficient and stronger visual saliency transformer. IEEE Transactions on Pattern Analysis and Machine Intelligence 46(11), 7300–7316 (2025)

    2025

  18. [26]

    Proceedings of the Advances in Neural Information Processing Systems37, 103031–103063 (2024)

    Liu, Y., Tian, Y., Zhao, Y., Yu, H., Xie, L., Wang, Y., Ye, Q., Jiao, J., Liu, Y.: Vmamba: Visual state space model. Proceedings of the Advances in Neural Information Processing Systems37, 103031–103063 (2024)

    2024

  19. [27]

    Liu, Z., Lin, Y., Cao, Y., Hu, H., Wei, Y., Zhang, Z., Lin, S., Guo, B.: Swin transformer:Hierarchical visiontransformerusingshiftedwindows.In:Proceedings of the IEEE/CVF International Conference on Computer Vision. pp. 10012–10022 (2021)

    2021

  20. [28]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition

    Liu, Z., Mao, H., Wu, C.Y., Feichtenhofer, C., Darrell, T., Xie, S.: A convnet for the 2020s. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. pp. 11976–11986 (2022)

    2022

  21. [29]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition

    Luo, Z., Liu, N., Zhao, W., Yang, X., Zhang, D., Fan, D.P., Khan, F., Han, J.: Vscode: General visual salient and camouflaged object detection with 2d prompt learning. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. pp. 17169–17180 (2024)

    2024

  22. [30]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition

    Lv, Y., Zhang, J., Dai, Y., Li, A., Liu, B., Barnes, N., Fan, D.P.: Simultane- ously 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

  23. [31]

    Lyu, P., Yu, X., Chi, J., Wu, H., Wu, C., Rajapakse, J.C.: Twinstnet: Broad-view twinstransformernetworkforbi-modalsalientobjectdetection.IEEETransactions on Image Processing (2025) 18 Duan et al

    2025

  24. [32]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition

    Mao, K., Wei, P., Lian, Y., Wang, Y., Zheng, N.: Beyond single-modal bound- ary: Cross-modal anomaly detection through visual prototype and harmonization. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. pp. 9964–9973 (2025)

    2025

  25. [33]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition

    Mei, H., Ji, G.P., Wei, Z., Yang, X., Wei, X., Fan, D.P.: Camouflaged object seg- mentation with distraction mining. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. pp. 8772–8781 (2021)

    2021

  26. [34]

    arXiv preprint arXiv:2304.07193 (2023)

    Oquab, M., Darcet, T., Moutakanni, T., Vo, 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

  27. [35]

    In: Proceedings of the AAAI Conference on Artificial Intelligence

    Pei, X., Huang, T., Xu, C.: Efficientvmamba: Atrous selective scan for light weight visual mamba. In: Proceedings of the AAAI Conference on Artificial Intelligence. vol. 39, pp. 6443–6451 (2025)

    2025

  28. [36]

    In: Proceedings of the European Conference on Computer Vision

    Peng, H., Li, B., Xiong, W., Hu, W., Ji, R.: Rgbd salient object detection: A bench- mark and algorithms. In: Proceedings of the European Conference on Computer Vision. pp. 92–109. Springer (2014)

    2014

  29. [37]

    In: Proceedings of the IEEE/CVF International Conference on Computer Vision

    Ranftl, R., Bochkovskiy, A., Koltun, V.: Vision transformers for dense prediction. In: Proceedings of the IEEE/CVF International Conference on Computer Vision. pp. 12179–12188 (2021)

    2021

  30. [38]

    In: Proceedings of the IEEE/CVF Interna- tional Conference on Computer Vision

    Ren, G., Liu, H., Lazarou, M., Stathaki, T.: Multi-modal segment anything model for camouflaged scene segmentation. In: Proceedings of the IEEE/CVF Interna- tional Conference on Computer Vision. pp. 19882–19892 (October 2025)

    2025

  31. [39]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition

    Rombach, R., Blattmann, A., Lorenz, D., Esser, P., Ommer, B.: High-resolution image synthesis with latent diffusion models. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. pp. 10684–10695 (2022)

    2022

  32. [40]

    Expert Systems with Applications275, 127004 (2025)

    Su, Y., Gao, H., Wang, M., Wang, F.: Asymmetric cross-modality interaction net- work for rgb-d salient object detection. Expert Systems with Applications275, 127004 (2025)

    2025

  33. [41]

    IEEE Transactions on Pattern Analysis and Machine Intelligence (2025)

    Sun, K., Chen, Z., Lin, X., Sun, X., Liu, H., Ji, R.: Conditional diffusion models for camouflaged and salient object detection. IEEE Transactions on Pattern Analysis and Machine Intelligence (2025)

    2025

  34. [42]

    In: Proceedings of the European Conference on Computer Vision

    Sun, Y., Xu, C., Yang, J., Xuan, H., Luo, L.: Frequency-spatial entanglement learn- ing for camouflaged object detection. In: Proceedings of the European Conference on Computer Vision. pp. 343–360 (2024)

    2024

  35. [43]

    IEEE Transactions on Circuits and Systems for Video Technology (2025)

    Wang, K., Tu, Z., Li, C., Liu, Z., Luo, B.: Unified-modal salient object detection via adaptive prompt learning. IEEE Transactions on Circuits and Systems for Video Technology (2025)

    2025

  36. [44]

    IEEE Transactions on Circuits and Systems for Video Technology34(8), 7344–7358 (2024)

    Wang, K., Tu, Z., Li, C., Zhang, C., Luo, B.: Learning adaptive fusion bank for multi-modal salient object detection. IEEE Transactions on Circuits and Systems for Video Technology34(8), 7344–7358 (2024)

    2024

  37. [45]

    In: Proceedings of the IEEE/CVF Confer- ence on Computer Vision and Pattern Recognition

    Wang, L., Yang, J., Zhang, Y., Wang, F., Zheng, F.: Depth-aware concealed crop detection in dense agricultural scenes. In: Proceedings of the IEEE/CVF Confer- ence on Computer Vision and Pattern Recognition. pp. 17201–17211 (2024)

    2024

  38. [48]

    In: Proceedings of the IEEE/CVF International Conference on Computer Vision

    Wang, W., Xie, E., Li, X., Fan, D.P., Song, K., Liang, D., Lu, T., Luo, P., Shao, L.: Pyramid vision transformer: A versatile backbone for dense prediction with- out convolutions. In: Proceedings of the IEEE/CVF International Conference on Computer Vision. pp. 568–578 (2021)

    2021

  39. [49]

    Wu, L., Gao, Z., Fei, H., Lee, M.L., Hsu, W.: Leaf-mamba: Local emphatic and adaptivefusionstatespacemodelforrgb-dsalientobjectdetection.In:Proceedings of the ACM International Conference on Multimedia (2025)

    2025

  40. [52]

    IEEE Signal Processing Letters29, 2577–2581 (2022)

    Xia, C., Duan, S., Ge, B., Zhang, H., Li, K.C.: Hdnet: Multi-modality hierarchy- aware decision network for rgb-d salient object detection. IEEE Signal Processing Letters29, 2577–2581 (2022)

    2022

  41. [53]

    Digital Signal Processing131, 103775 (2022)

    Xia, C., Duan, S., Fang, X., Gao, X., Sun, Y., Ge, B., Zhang, H., Li, K.C.: Efgnet: Encoder steered multi-modality feature guidance network for rgb-d salient object detection. Digital Signal Processing131, 103775 (2022)

    2022

  42. [54]

    Proceedings of the Advances in Neural Information Processing Systems35, 4203–4217 (2022)

    Yang, J., Li, C., Dai, X., Gao, J.: Focal modulation networks. Proceedings of the Advances in Neural Information Processing Systems35, 4203–4217 (2022)

    2022

  43. [55]

    IEEE Transactions on Circuits and Systems for Video Technology34(3), 1481–1492 (2023)

    Yang, J., Bai, L., Sun, Y., Tian, C., Mao, M., Wang, G.: Pixel difference convo- lutional network for rgb-d semantic segmentation. IEEE Transactions on Circuits and Systems for Video Technology34(3), 1481–1492 (2023)

    2023

  44. [56]

    Proceedings of the Advances in Neural Information Processing Systems37, 21875–21911 (2024)

    Yang, L., Kang, B., Huang, Z., Zhao, Z., Xu, X., Feng, J., Zhao, H.: Depth anything v2. Proceedings of the Advances in Neural Information Processing Systems37, 21875–21911 (2024)

    2024

  45. [57]

    IEEE Transactions on Pattern Analysis and Machine Intelligence (2024)

    Yin, B., Zhang, X., Fan, D.P., Jiao, S., Cheng, M.M., Van Gool, L., Hou, Q.: Camoformer: Masked separable attention for camouflaged object detection. IEEE Transactions on Pattern Analysis and Machine Intelligence (2024)

    2024

  46. [58]

    In: Proceedings of the ACM International Conference on Multimedia

    Yu, Z., Zhang, X., Zhao, L., Bin, Y., Xiao, G.: Exploring deeper! segment anything model with depth perception for camouflaged object detection. In: Proceedings of the ACM International Conference on Multimedia. pp. 4322–4330 (2024)

    2024

  47. [59]

    In: Proceedings of the European Conference on Computer Vision

    Yuan, H., Li, X., Yang, Y., Cheng, G., Zhang, J., Tong, Y., Zhang, L., Tao, D.: Polyphonicformer: Unified query learning for depth-aware video panoptic segmen- tation. In: Proceedings of the European Conference on Computer Vision. pp. 582– 599 (2022)

    2022

  48. [60]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition

    Zhai, Q., Li, X., Yang, F., Chen, C., Cheng, H., Fan, D.P.: Mutual graph learning for camouflaged object detection. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. pp. 12997–13007 (2021)

    2021

  49. [61]

    In: Proceedings of the ACM International Conference on Multimedia (2025)

    Zhang, C., Zhang, Q., Wu, J., Pang, Y.: Cgcod: Class-guided camouflaged object detection. In: Proceedings of the ACM International Conference on Multimedia (2025)

    2025

  50. [64]

    In: Proceedings of the AAAI Conference on Artificial Intelligence

    Zhang, S., Huang, J., Tang, W., Wu, Y., Hu, T., Xu, X., Liu, J.: Dimsod: A diffusion-based framework for multi-modal salient object detection. In: Proceedings of the AAAI Conference on Artificial Intelligence. vol. 39, pp. 10103–10111 (2025)

    2025

  51. [65]

    In: Proceedings of the European Conference on Computer Vision

    Zhao, J., Li, X., Yang, F., Zhai, Q., Luo, A., Jiao, Z., Cheng, H.: Focusdiffuser: Perceiving local disparities for camouflaged object detection. In: Proceedings of the European Conference on Computer Vision. pp. 181–198 (2024)

    2024

  52. [66]

    In: Proceedings of the IEEE/CVF International Confer- ence on Computer Vision

    Zhou, T., Fu, H., Chen, G., Zhou, Y., Fan, D.P., Shao, L.: Specificity-preserving rgb-d saliency detection. In: Proceedings of the IEEE/CVF International Confer- ence on Computer Vision. pp. 4681–4691 (2021)

    2021

  53. [67]

    IEEE Transactions on Image Processing31, 7036–7047 (2022)

    Zhou, T., Zhou, Y., Gong, C., Yang, J., Zhang, Y.: Feature aggregation and prop- agation network for camouflaged object detection. IEEE Transactions on Image Processing31, 7036–7047 (2022)

    2022

  54. [68]

    In: Proceedings of the IEEE/CVF International Conference on Computer Vision

    Zhou, Z., Li, Y., Zhong, C., Huang, J., Pei, J., Li, H., Tang, H.: Rethinking de- tecting salient and camouflaged objects in unconstrained scenes. In: Proceedings of the IEEE/CVF International Conference on Computer Vision. pp. 22372–22382 (October 2025)

    2025

  55. [69]

    arXiv preprint arXiv:2401.09417 (2024)

    Zhu, L., Liao, B., Zhang, Q., Wang, X., Liu, W., Wang, X.: Vision mamba: Efficient visual representation learning with bidirectional state space model. arXiv preprint arXiv:2401.09417 (2024)

    2024 arXiv

  56. [70]

    Pattern Recognition155, 110543 (2024)

    Zhu, X.F., Xu, T., Atito, S., Awais, M., Wu, X.J., Feng, Z., Kittler, J.: Self- supervised learning for rgb-d object tracking. Pattern Recognition155, 110543 (2024)

    2024

  57. [71]

    In: Pro- ceedings of the AAAI Conference on Artificial Intelligence

    Zhu, X.F., Xu, T., Tang, Z., Wu, Z., Liu, H., Yang, X., Wu, X.J., Kittler, J.: Rgbd1k: A large-scale dataset and benchmark for rgb-d object tracking. In: Pro- ceedings of the AAAI Conference on Artificial Intelligence. vol. 37, pp. 3870–3878 (2023)

    2023

Pith tools

Reviewed August 1, 2026 · model on record in the stance chip above.