Pith. sign in

REVIEW 4 major objections 7 minor 60 references

Stronger, Steadier & Superior: Geometric Consistency in Depth VFM Forges Domain Generalized Semantic Segmentation

T0 review · 4 major / 7 minor · reviewed 2026-08-16 · deepseek-v4-flash

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

Pith's one-line read DepthForge claims that fusing frozen depth features into frozen visual foundation models, through depth-aware learnable tokens, yields domain-generalized semantic segmentation that is stronger in normal scenes and dramatically better…

desk verdict Useful, incremental DGSS paper combining frozen visual and depth VFMs with learnable tokens; the +4% headline claim is inflated and depth quality on target domains is unverified, but the core recipe holds up. read the letter →

arxiv 2504.12753 v3 pith:TCN7UWDL submitted 2025-04-17 cs.CV

Siyu Chen , Ting Han , Changshe Zhang , Xin Luo , Meiliu Wu , Guorong Cai , Jinhe Su This is my paper · ORCID
classification cs.CV
keywords domaingeneralizedsemanticsegmentationvisionfoundationmodelsdepth-awarelearnabletokensgeometricconsistencymulti-modalfusionadverseconditionsparameter-efficientfine-tuningDINOv2
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

DepthForge sets out to prove that geometry, not pixels, is the ingredient that lets a segmentation model survive a domain shift. The paper argues that RGB cues are the first thing to fail in night, snow, fog, or rain, while relative depth stays roughly stable, so it builds a fine-tuning strategy that injects depth cues from a frozen Depth Anything V2 model into the frozen features of DINOv2 or EVA02 at every layer. The injection happens through depth-aware learnable tokens and an attention-bias term that adds a depth-based spatial relationship to the visual attention map. On the standard DGSS benchmarks, the method reports consistent gains over prior VFM-based methods, with the largest jumps under extreme conditions: 75.4 vs. 70.6 mIoU on snow and 60.4 vs. 55.9 at night in the Cityscapes-to-ACDC setting. If the claim holds, the practical payoff is that reliable segmentation under adverse weather can be obtained by reusing frozen foundation models plus a small set of tunable tokens, rather than retraining on target-domain data.

What carries the argument

The load-bearing mechanism is the depth-aware learnable token $T_i \in \mathbb{R}^{m\times c}$ and the attention rule $A_i = \mathrm{Softmax}(f^v_i (T^v_i)^\top / \sqrt{c}) + \lambda \mathrm{Softmax}(f^d_i (T^d_i)^\top / \sqrt{c})$, which adds a depth-derived spatial bias to the visual attention map so that pixels are tied to real-world locations even when RGB features are ambiguous. The layer update $f_{i+1} = W^v_i f_i + \varepsilon(\epsilon_v(T_i(W^v_i f^v_i)) + \epsilon_d(T_i(W^d_i f^d_i)))$ lets each modality contribute independently through the same token set, preventing one cue from dominating. Around these two identities, DepthForge wraps an Attention Optimization module that rescales and gates the attention output, and a Depth Refinement Decoder that fuses multi-layer enhanced features through MLPs, concatenation, a convolution, and Transformer layers. The depth-aware token is what carries the paper's claim: it is the place where visual and depth cues are decoupled and recombined, and where the geometric-consistency bias enters the frozen VFM at every layer.

What would settle it

Compare DepthForge's mIoU on the Cityscapes to ACDC night and snow splits when the frozen depth features are replaced at test time by corrupted versions, such as Gaussian noise or a constant plane. If the segmentation barely degrades, the claimed geometric-consistency mechanism is not what produces the gain; if it degrades sharply, the gain is tied to depth quality as argued. A complementary check is to compute Depth Anything V2's depth error on ACDC night and snow images and see whether image-level gains correlate with depth accuracy.

Watch

Extended reading notes

Core claim

The central claim is that a frozen depth vision foundation model can be used as a stable geometric prior for domain-generalized semantic segmentation, and that the correct way to combine it with a frozen visual VFM is per-layer, through learnable tokens that carry both visual and depth awareness. DepthForge updates the layer feature not by adding depth features directly, but by generating an enhanced feature from visual and depth tokens via Eq. (3), and it forges geometric consistency by adding a depth-attention bias to the visual attention map in Eq. (5). The paper shows that this combination outperforms simpler fusion strategies (concatenation, depth-prompts, direct token injection) and that the gains concentrate precisely where visual cues are weakest: snow, night, fog, and rain. The authors argue the resulting visual-spatial attention is steadier across domains, and they support that with t-SNE and affinity-map visualizations showing more separated class clusters and cleaner attention maps.

Load-bearing premise

The method's edge depends on the depth model producing trustworthy relative depth on the unseen target domains: if Depth Anything V2's estimates are noisy or wrong under night, fog, and snow, the spatial-consistency bias becomes noise and the advantage should shrink or vanish. The paper never measures depth quality on those target domains, so this premise is asserted rather than verified.

Editorial extensions

If this is right

  • On Cityscapes to ACDC, DepthForge reports the best results among VFM-based DGSS methods on all four adverse conditions, with the largest margins on snow and night (+4.8 and +4.5 mIoU over REIN).
  • Across GTA5 to Cityscapes+BDD+Mapillary and Cityscapes to BDD+Mapillary+GTA5, DepthForge improves mIoU over REIN by roughly 2.4 to 3.1 points and over FADA by about 0.1 to 1.1 points while using the same or fewer trainable parameters.
  • The component ablations attribute the gain to the full DepthForge configuration (depth-aware tokens plus attention optimization plus refinement decoder); simply concatenating depth features or adding depth prompts gives little or no improvement.
  • The gains hold when the visual backbone is changed to EVA02, CLIP, or SAM, and relative-depth cues from Depth Anything V2 beat absolute-depth prompts from Prompt Depth Anything.
  • The per-layer token design means the same frozen VFMs can be reused for different segmentation tasks with only the small tunable token set and the head retrained.

Reading between the lines

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

  • Beyond the paper: because the design leans on relative depth rather than absolute distance, any monocular depth model with reliable ordinal relationships (other depth VFMs, or even stereo or disparity networks) should be able to substitute for Depth Anything V2, which would make the recipe model-agnostic.
  • Beyond the paper: the same depth-aware token fusion could be transferred to other dense prediction tasks that suffer under domain shift, such as monocular depth estimation itself or panoptic and instance segmentation, since the mechanism only modifies how frozen features are combined, not the task head.
  • Beyond the paper: a natural stress test is to degrade or zero out the depth branch for a subset of target images; if the paper's story is right, the model's attention maps should become visibly less stable exactly on those images, linking the qualitative steadier-attention claim to a quantitative input.
  • Beyond the paper: because the authors never measure target-domain depth quality, the most useful follow-up is a calibration study that gates the depth bias $\lambda$ when the depth model is uncertain; if the gate recovers performance when depth is poor, the approach becomes more robust than the fixed-bias version.
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 / 7 minor

Summary. The paper proposes DepthForge, a fine-tuning framework for domain-generalized semantic segmentation (DGSS) that combines features from a frozen visual VFM (DINOv2 or EVA02) with features from a frozen depth VFM (Depth Anything V2). The method introduces depth-aware learnable tokens, an attention-optimization module, and a multi-scale depth refinement decoder, and it is evaluated under three DGSS protocols over Cityscapes, ACDC, BDD100k, Mapillary, and GTA5. The experiments report consistent mIoU gains over REIN and FADA across all settings, with the largest gains in extreme conditions, e.g., 75.4 vs. 70.6 on Snow and 60.4 vs. 55.9 on Night under Cityscapes-to-ACDC (Table 1). The abstract and introduction additionally claim an approximate +4% mIoU improvement in the GTA5-to-Cityscapes+BDD100k+Mapillary setting and roughly +5% in extreme scenes.

Significance. If the reported gains hold up, DepthForge is a useful contribution: it achieves consistent improvements over strong VFM-based DGSS baselines across three protocols and two visual backbones at constant parameter cost (2.99M trainable parameters, equal to REIN and well below FADA's 11.65M, Tables 6-7), and the comparison set is thorough, including ResNet, Transformer, and VFM baselines as well as PEFT methods such as LoRA, VPT, and AdaptFormer (Table 7). The direction of using frozen depth-VFM features as a geometric prior for DGSS is timely and the code is promised. The main caveats are attribution and reproducibility: the headline +4% is not supported by Table 3, the ablation record is internally inconsistent (Tables 4 and 5 report different numbers for the same configuration), and a key fusion parameter in Eq. (5) is never specified. These issues are fixable within the scope of the manuscript but need to be resolved before the geometric-consistency mechanism can be accepted as the cause of the gains.

major comments (4)
  1. [Abstract, Section 1, Table 3] The abstract and the introduction claim approximately +4% mIoU improvement in the GTA5-to-Cityscapes+BDD100k+Mapillary setting, but Table 3 shows average gains of 2.73 points over REIN (2.64, 2.42, and 3.12 per domain) and under 1.1 points over FADA; no single domain reaches +4%. The accompanying '+5% in extreme scenes' claim refers to the Cityscapes-to-ACDC setting in Table 1 (Snow +4.8, Night +4.5 over REIN), not to the GTA5 setting, so the introduction conflates two evaluation protocols, and it does not state which baseline the percentages are measured against. These claims should be revised to match the tables.
  2. [Table 4, Section 5.4] The ablation table does not state whether its rows are cumulative additions to REIN or independent configurations. Under the natural cumulative reading, the final increment (+DA+AO to DepthForge) contributes 3.2 points on Snow and 3.0 points on Night, which are the two headline extreme-condition gains, whereas the depth-awareness module (+DA+AO over REIN) contributes only 1.5-1.7 points on those conditions. This is at odds with Section 5.4, which attributes 'an approximate 2% improvement' primarily to the depth-aware tokens and never discusses the decoder's dominant contribution to the extreme-condition gains. Additionally, the same configuration is reported with different numbers in Tables 4 and 5: '+concat f_d^i' is 70.8/56.1/79.4/72.8 in Table 4 but appears as 69.8/55.0/78.3/71.2 under 'Concatenation' in Table 5. The row semantics should be stated explicitly, the two tables reconciled, and the decomposition of the extreme-condition gains discussed.
  3. [Eq. (5), Section 4.2, Table 5, Section 5.2] The depth-attention weight lambda in Eq. (5) and the gating threshold for 'discarding high-weight features' in Section 4.2 are never assigned values or selection procedures in Section 5.2. Table 5 shows that removing the scale factor changes the average gain from 3.5 to 2.8 points, so this parameter is material to the result. Without reporting its value and the threshold, the central fusion mechanism is not reproducible, and one cannot rule out that these constants were selected on the target test domains. Please specify both, and state whether any hyperparameter was chosen using validation performance on the unseen target sets.
  4. [Sections 4 and 5.3] The paper's mechanism claim is that Depth Anything V2 supplies geometric consistency because 'the underlying geometry remains stable' across domains while visual cues fail, but no evidence is provided that the depth features are actually accurate or informative on the unseen target domains, particularly night and snow, where monocular depth estimates are known to degrade. The construction also assumes that the layer indices of the frozen visual and depth encoders align semantically, which is asserted without justification. Please add direct evidence, such as depth maps on target-domain samples, a quantitative depth-quality measure, or an ablation that perturbs or corrupts the depth features to verify sensitivity. Without this, the geometric-consistency attribution rests on an unverified premise.
minor comments (7)
  1. [Eq. (4)] The notation in Eq. (4) is garbled: 'Hθ⟨(FΘv,Θd,θD(xi)),yi' should be a standard loss over the segmentation head, and the parameters θ_h and Hθ are used inconsistently.
  2. [Table 5] The column header 'Rain Night Snow Rog' is scrambled and mistyped; the numbers are in Snow/Night/Fog/Rain order, and 'Rog' should be 'Fog'. The row labels (No Depth vs. REIN, Depth Token vs. the rows of Table 4) should also be harmonized with Table 4.
  3. [Section 5.3, Table 2] The text reports GTA gains of 3.12% and 1.13% over REIN and FADA under the Cityscapes-to-BDD+Map+GTA setting, but Table 2 implies gains of 4.83 and 3.46 points (67.24 vs. 62.41 and 63.78); the printed numbers appear to be copied from Table 3 and should be corrected.
  4. [Tables 1-3] No variance or multiple-seed information is reported, and several gains over the second-best method are at or below one point (e.g., 0.07 on Mapillary in Table 2 and 1.0 on Fog in Table 1), so the claim of significant improvement needs error bars or repeated runs to be meaningful.
  5. [Figure 2, Table 5] The Config.1/Config.2/Config.3 definitions in the Figure 2 caption should be aligned with the Config1/Config2/Config3 labels in Table 5 to avoid confusion about which configuration is which.
  6. [Throughout] There are several typos and inconsistencies: 'datsets' in the Abstract, 'optimaiton' in Section 4, duplicated references [39] and [46]/[47], inconsistent GitHub URLs between the abstract (anonymouse-xzrptkvyqc) and the body (SY-Ch), and 'qualitatively evaluating' in Section 4.1 should be 'quantitatively evaluating'.
  7. [Section 1, Section 2.2] The claim of being 'the first depth-aware framework specifically designed for DGSS' is too strong given that Section 2.2 cites existing RGB-D domain-generalization works [19, 27, 38, 55]; the novelty claim should be scoped relative to that literature.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: DepthForge is an empirical fusion design with independently ablated components; the claimed gains do not reduce by construction to the method's inputs.

full rationale

DepthForge proposes a depth-aware fine-tuning framework whose core equations (Eq. 3, Eq. 5, Eq. 6) combine frozen visual features, frozen depth features, and learnable tokens, and it validates each component with ablations (Tables 4 and 5). The paper's inputs are DINOv2/EVA02 visual features and Depth Anything V2 depth features; its outputs are mIoU improvements on unseen target domains. No equation equates the claimed improvement with a fitted parameter or with the input features by construction: Eq. (5) is an attention sum with learnable tokens and a scale factor, Eq. (6) is a token-weighted refinement, and the decoder (Eqs. 7-8) fuses multi-layer features. The ablations independently compare configurations (Config. 1, Config. 2, Config. 3 in Table 5), so the reported gains are not equivalent to training-fit values. The self-citations (references [8], [9], [17]) appear only in related-work discussion of RGB-D segmentation and are not load-bearing for the central claim. The unverified depth-quality premise on night and snow targets and the unexplained gap between the '+DA+AO' row and the final 'DepthForge' row in Table 4 are correctness and consistency concerns, not circularity under the stated criteria. This is an empirical benchmark study rather than a derivation, and no prediction reduces to its own input by construction.

Assumptions & free parameters 3 free parameters · 3 assumptions · 1 invented entities

The central claim rests on a set of empirical design choices and assumptions about cross-modal layer alignment and depth robustness, plus two hyperparameters (lambda and scale factor) that are not reported. These are the main unverified inputs beyond the trained network weights.

free parameters (3)
  • lambda (depth attention weight) = not reported
    Introduced in Eq. (5) to weight the depth attention bias; no value or ablation is given, and it likely was tuned on the target domains.
  • scale factor in DepthForge fusion = not reported
    Referenced in Table 5 as 'w/o Scale Factor' showing a performance drop, but the value is not specified in the text.
  • gating threshold for high-weight feature discarding = not reported
    Mentioned in Section 4.2 as 'discarding high-weight features when the features are sufficiently precise'; the condition is not quantified.
assumptions (3)
  • ad hoc to paper DINOv2 or EVA02 and Depth Anything V2 layer indices align semantically across the two frozen backbones.
    The fusion in Eq. (3) assumes layer i of the visual VFM corresponds to layer i of the depth VFM; no justification is given for this cross-modal layer correspondence.
  • domain assumption Depth Anything V2 produces reliable depth estimates on the unseen target domains, including night, snow, fog, and rain.
    The method's benefit under extreme conditions depends on the depth prior being accurate on those same conditions, but the paper does not evaluate depth quality on the target domains.
  • standard math Adding two softmax attention maps in Eq. (5) preserves a reasonable attention distribution.
    The equation adds the visual and depth attention maps without renormalization, implicitly assuming compatible scales and that the sum behaves like an attention map.
invented entities (1)
  • Depth-aware learnable tokens independent evidence
    purpose: Per-layer tokens that ingest both visual and depth cues to refine frozen VFM features with spatial structure.
    Ablation studies in Tables 4 and 5 show mIoU gains when these tokens are added, providing empirical evidence beyond the conceptual proposal.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Stronger, Steadier & Superior: Geometric Consistency in Depth VFM Forges Domain Generalized Semantic Segmentation." pith.science (2026). https://pith.science/paper/TCN7UWDL

@misc{pith2026250412753,
  author       = {Pith},
  title        = {Pith review of: Stronger, Steadier & Superior: Geometric Consistency in Depth VFM Forges Domain Generalized Semantic Segmentation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/TCN7UWDL}},
  note         = {Machine review of arXiv:2504.12753}
}
read the original abstract

Vision Foundation Models (VFMs) have delivered remarkable performance in Domain Generalized Semantic Segmentation (DGSS). However, recent methods often overlook the fact that visual cues are susceptible, whereas the underlying geometry remains stable, rendering depth information more robust. In this paper, we investigate the potential of integrating depth information with features from VFMs, to improve the geometric consistency within an image and boost the generalization performance of VFMs. We propose a novel fine-tuning DGSS framework, named DepthForge, which integrates the visual cues from frozen DINOv2 or EVA02 and depth cues from frozen Depth Anything V2. In each layer of the VFMs, we incorporate depth-aware learnable tokens to continuously decouple domain-invariant visual and spatial information, thereby enhancing depth awareness and attention of the VFMs. Finally, we develop a depth refinement decoder and integrate it into the model architecture to adaptively refine multi-layer VFM features and depth-aware learnable tokens. Extensive experiments are conducted based on various DGSS settings and five different datsets as unseen target domains. The qualitative and quantitative results demonstrate that our method significantly outperforms alternative approaches with stronger performance, steadier visual-spatial attention, and superior generalization ability. In particular, DepthForge exhibits outstanding performance under extreme conditions (e.g., night and snow). Code is available at https://github.com/anonymouse-xzrptkvyqc/DepthForge.

Figures

Figures reproduced from arXiv: 2504.12753 by the authors.

Figure 1
Figure 1. Existing methods fail to fully exploit the potential of VFMs for DGSS when visual cues are limited or absent. To this end, [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Different configurations of incorporating depth cues into [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗
Figure 3
Figure 3. DepthForge framework. It incorporates the features from frozen [PITH_FULL_IMAGE:figures/full_fig_p003_3.png] view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: Key segmentation examples of existing DGSS methods and DepthForge under the Cityscapes to ACDC unseen target domains [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]
Figure 5
Figure 5. Figure 5: Key segmentation examples of existing DGSS methods and DepthForge under the GTA5 to Mapillary, BDD100k, and Cityscapes [PITH_FULL_IMAGE:figures/full_fig_p005_5.png]
Figure 6
Figure 6. Figure 6: t-SNE visualization compares REIN and our [PITH_FULL_IMAGE:figures/full_fig_p006_6.png]
Figure 7
Figure 7. Figure 7: Visualizations of attention maps and affinity maps under differnet configurations, where DRD denotes the Depth Refinement [PITH_FULL_IMAGE:figures/full_fig_p007_7.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

60 extracted references · 52 canonical work pages

  1. [1]

    Style blind domain generalized semantic segmentation via covariance alignment and semantic con- sistence contrastive learning

    Woo-Jin Ahn, Geun-Yeong Yang, Hyun-Duck Choi, and Myo-Taeg Lim. Style blind domain generalized semantic segmentation via covariance alignment and semantic con- sistence contrastive learning. In CVPR, pages 3616–3626,

    work page

  2. [2]

    Learning frequency- adapted vision foundation model for domain generalized se- mantic segmentation

    Qi Bi, Jingjun Yi, Hao Zheng, Haolan Zhan, Yawen Huang, Wei Ji, Yuexiang Li, and Yefeng Zheng. Learning frequency- adapted vision foundation model for domain generalized se- mantic segmentation. NeurIPS, 37:94047–94072, 2024. 1, 2, 4, 6, 8

    work page 2024

  3. [3]

    Learning content- enhanced mask transformer for domain generalized urban- scene segmentation

    Qi Bi, Shaodi You, and Theo Gevers. Learning content- enhanced mask transformer for domain generalized urban- scene segmentation. In AAAI, pages 819–827, 2024. 4, 6

    work page 2024

  4. [4]

    Ex- ploiting domain-specific features to enhance domain gener- alization

    Manh-Ha Bui, Toan Tran, Anh Tran, and Dinh Phung. Ex- ploiting domain-specific features to enhance domain gener- alization. In NeurIPS, pages 21189–21201. Curran Asso- ciates, Inc., 2021. 2

    work page 2021

  5. [5]

    Pasta: Proportional amplitude spectrum training augmentation for syn-to-real domain gen- eralization

    Prithvijit Chattopadhyay, Kartik Sarangmath, Vivek Vi- jaykumar, and Judy Hoffman. Pasta: Proportional amplitude spectrum training augmentation for syn-to-real domain gen- eralization. In ICCV, pages 19288–19300, 2023. 8

    work page 2023

  6. [6]

    Em-trans: Edge-aware multi- modal transformer for rgb-d salient object detection

    Geng Chen, Qingyue Wang, Bo Dong, Ruitao Ma, Nian Liu, Huazhu Fu, and Yong Xia. Em-trans: Edge-aware multi- modal transformer for rgb-d salient object detection. IEEE Transactions on Neural Networks and Learning Systems, 36 (2):3175–3188, 2024. 3

    work page 2024

  7. [7]

    Adaptformer: Adapting vision transformers for scalable visual recognition

    Shoufa Chen, Chongjian Ge, Zhan Tong, Jiangliu Wang, Yibing Song, Jue Wang, and Ping Luo. Adaptformer: Adapting vision transformers for scalable visual recognition. NeurIPS, 35:16664–16678, 2022. 8

    work page 2022

  8. [8]

    Depth Matters: Exploring Deep Interactions of RGB-D for Semantic Segmentation in Traffic Scenes

    Siyu Chen, Ting Han, Changshe Zhang, Weiquan Liu, Jinhe Su, Zongyue Wang, and Guorong Cai. Depth matters: Ex- ploring deep interactions of rgb-d for semantic segmentation in traffic scenes. arXiv preprint arXiv:2409.07995, 2024. 3

    work page Pith review arXiv 2024

Show all 60 references
  1. [9]

    Hspformer: Hierarchical spatial perception transformer for semantic segmentation

    Siyu Chen, Ting Han, Changshe Zhang, Jinhe Su, Ruisheng Wang, Yiping Chen, Zongyue Wang, and Guorong Cai. Hspformer: Hierarchical spatial perception transformer for semantic segmentation. IEEE Transactions on Intelligent Transportation Systems, 2025. 3

    2025

  2. [10]

    Robustnet: Improving domain generalization in urban-scene segmentation via in- stance selective whitening

    Sungha Choi, Sanghun Jung, Huiwon Yun, Joanne T Kim, Seungryong Kim, and Jaegul Choo. Robustnet: Improving domain generalization in urban-scene segmentation via in- stance selective whitening. In CVPR, pages 11580–11590,

  3. [11]

    MMSegmentation: Openmmlab semantic segmentation toolbox and benchmark

    MMSegmentation Contributors. MMSegmentation: Openmmlab semantic segmentation toolbox and benchmark. https : / / github . com / open - mmlab/mmsegmentation, 2020. 6

    2020

  4. [12]

    The cityscapes dataset for semantic urban scene understanding

    Marius Cordts, Mohamed Omran, Sebastian Ramos, Timo Rehfeld, Markus Enzweiler, Rodrigo Benenson, Uwe Franke, Stefan Roth, and Bernt Schiele. The cityscapes dataset for semantic urban scene understanding. In CVPR, pages 3213–3223, 2016. 6

    2016

  5. [13]

    Hgformer: Hierarchical grouping transformer for domain generalized semantic segmentation

    Jian Ding, Nan Xue, Gui-Song Xia, Bernt Schiele, and Dengxin Dai. Hgformer: Hierarchical grouping transformer for domain generalized semantic segmentation. In CVPR, pages 15413–15423, 2023. 4

    2023

  6. [14]

    Asymformer: Asymmetrical cross-modal representation learning for mobile platform real-time rgb-d semantic segmentation

    Siqi Du, Weixi Wang, Renzhong Guo, Ruisheng Wang, and Shengjun Tang. Asymformer: Asymmetrical cross-modal representation learning for mobile platform real-time rgb-d semantic segmentation. In CVPRW, pages 7608–7615, 2024. 3

    2024

  7. [15]

    Eva-02: A visual representation for neon genesis

    Yuxin Fang, Quan Sun, Xinggang Wang, Tiejun Huang, Xin- long Wang, and Yue Cao. Eva-02: A visual representation for neon genesis. arXiv preprint arXiv:2303.11331, 2023. 1, 2, 7, 8

    2023 arXiv

  8. [16]

    Eva: Exploring the limits of masked visual representa- tion learning at scale

    Yuxin Fang, Wen Wang, Binhui Xie, Quan Sun, Ledell Wu, Xinggang Wang, Tiejun Huang, Xinlong Wang, and Yue Cao. Eva: Exploring the limits of masked visual representa- tion learning at scale. In CVPR, pages 19358–19369, 2023. 7, 8

    2023

  9. [17]

    Epurate-net: Efficient progressive uncertainty refinement analysis for traffic envi- ronment urban road detection

    Ting Han, Siyu Chen, Chuanmu Li, Zongyue Wang, Jinhe Su, Min Huang, and Guorong Cai. Epurate-net: Efficient progressive uncertainty refinement analysis for traffic envi- ronment urban road detection. IEEE Transactions on Intelli- gent Transportation Systems, 2024. 3

    2024

  10. [18]

    Unsupervised do- main generalization by learning a bridge across domains

    Sivan Harary, Eli Schwartz, Assaf Arbelle, Peter Staar, Shady Abu-Hussein, Elad Amrani, Roei Herzig, Amit Al- fassy, Raja Giryes, Hilde Kuehne, Dina Katabi, Kate Saenko, Rogerio Feris, and Leonid Karlinsky. Unsupervised do- main generalization by learning a bridge across domai...

    2022

  11. [19]

    Learning depth repre- sentation from rgb-d videos by time-aware contrastive pre- training

    Zongtao He, Liuyi Wang, Ronghao Dang, Shu Li, Qingqing Yan, Chengju Liu, and Qijun Chen. Learning depth repre- sentation from rgb-d videos by time-aware contrastive pre- training. IEEE TCSVT, 34(6):4143–4158, 2024. 3

    2024

  12. [20]

    Lora: Low-rank adaptation of large language models

    Edward J Hu, Yelong Shen, Phillip Wallis, Zeyuan Allen- Zhu, Yuanzhi Li, Shean Wang, Lu Wang, and Weizhu Chen. Lora: Low-rank adaptation of large language models. arXiv preprint arXiv:2106.09685, 2021. 8

    2021 arXiv

  13. [21]

    Itera- tive normalization: Beyond standardization towards efficient whitening

    Lei Huang, Yi Zhou, Fan Zhu, Li Liu, and Ling Shao. Itera- tive normalization: Beyond standardization towards efficient whitening. In CVPR, pages 4874–4883, 2019. 4

    2019

  14. [22]

    Style pro- jected clustering for domain generalized semantic segmenta- tion

    Wei Huang, Chang Chen, Yong Li, Jiacheng Li, Cheng Li, Fenglong Song, Youliang Yan, and Zhiwei Xiong. Style pro- jected clustering for domain generalized semantic segmenta- tion. In CVPR, pages 3061–3071, 2023. 6

    2023

  15. [23]

    Vi- sual prompt tuning

    Menglin Jia, Luming Tang, Bor-Chun Chen, Claire Cardie, Serge Belongie, Bharath Hariharan, and Ser-Nam Lim. Vi- sual prompt tuning. In ECCV, pages 709–727. Springer,

  16. [24]

    Pin the memory: Learning to generalize semantic segmentation

    Jin Kim, Jiyoung Lee, Jungin Park, Dongbo Min, and Kwanghoon Sohn. Pin the memory: Learning to generalize semantic segmentation. In CVPR, 2022. 1 9

    2022

  17. [25]

    Segment any- thing

    Alexander Kirillov, Eric Mintun, Nikhila Ravi, Hanzi Mao, Chloe Rolland, Laura Gustafson, Tete Xiao, Spencer White- head, Alexander C Berg, Wan-Yen Lo, et al. Segment any- thing. In ICCV, pages 4015–4026, 2023. 1, 7

    2023

  18. [26]

    Wildnet: Learning domain generalized semantic seg- mentation from the wild

    Suhyeon Lee, Hongje Seong, Seongwon Lee, and Euntai Kim. Wildnet: Learning domain generalized semantic seg- mentation from the wild. In CVPR, pages 9936–9946, 2022. 4, 6

    2022

  19. [27]

    Deep hypersphere feature regularization for weakly supervised rgb-d salient object detection

    Zhiyu Liu, Munawar Hayat, Hong Yang, Duo Peng, and Yin- jie Lei. Deep hypersphere feature regularization for weakly supervised rgb-d salient object detection. IEEE TIP , 32: 5423–5437, 2023. 3

    2023

  20. [28]

    The mapillary vistas dataset for semantic understanding of street scenes

    Gerhard Neuhold, Tobias Ollmann, Samuel Rota Bulo, and Peter Kontschieder. The mapillary vistas dataset for semantic understanding of street scenes. In ICCV, pages 4990–4999,

  21. [29]

    Generalization by adaptation: Diffusion-based domain extension for domain-generalized semantic segmentation

    Joshua Niemeijer, Manuel Schwonberg, Jan-Aike Term¨ohlen, Nico M Schmidt, and Tim Fingscheidt. Generalization by adaptation: Diffusion-based domain extension for domain-generalized semantic segmentation. In Proceedings of the IEEE/CVF Winter Conference on Applications of Compu...

    2024

  22. [30]

    Dinov2: Learning robust visual features without supervision

    Maxime Oquab, Timoth ´ee Darcet, Th ´eo Moutakanni, Huy V o, Marc Szafraniec, Vasil Khalidov, Pierre Fernandez, Daniel Haziza, Francisco Massa, Alaaeldin El-Nouby, et al. Dinov2: Learning robust visual features without supervision. arXiv preprint arXiv:2304.07193, 2023. 2, 7, 8

    2023 arXiv

  23. [31]

    Two at once: Enhancing learning and generalization capacities via ibn-net

    Xingang Pan, Ping Luo, Jianping Shi, and Xiaoou Tang. Two at once: Enhancing learning and generalization capacities via ibn-net. In ECCV, pages 464–479, 2018. 4

    2018

  24. [32]

    Switchable whitening for deep representation learning

    Xingang Pan, Xiaohang Zhan, Jianping Shi, Xiaoou Tang, and Ping Luo. Switchable whitening for deep representation learning. In CVPR, pages 1863–1871, 2019. 4

    2019

  25. [33]

    Global and local texture randomization for synthetic-to-real semantic segmentation

    Duo Peng, Yinjie Lei, Lingqiao Liu, Pingping Zhang, and Jun Liu. Global and local texture randomization for synthetic-to-real semantic segmentation. IEEE TIP , 30: 6594–6608, 2021. 4, 6, 8

    2021

  26. [34]

    Semantic-aware domain generalized segmentation

    Duo Peng, Yinjie Lei, Munawar Hayat, Yulan Guo, and Wen Li. Semantic-aware domain generalized segmentation. In CVPR, pages 2584–2595, 2022. 1

    2022

  27. [35]

    Semantic-aware domain generalized segmentation

    Duo Peng, Yinjie Lei, Munawar Hayat, Yulan Guo, and Wen Li. Semantic-aware domain generalized segmentation. In CVPR, pages 2594–2605, 2022. 4, 6

    2022

  28. [36]

    Learning transferable visual models from natural language supervi- sion

    Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, et al. Learning transferable visual models from natural language supervi- sion. In International conference on machine learning, ...

    2021

  29. [37]

    Playing for data: Ground truth from computer games

    Stephan R Richter, Vibhav Vineet, Stefan Roth, and Vladlen Koltun. Playing for data: Ground truth from computer games. In ECCV, pages 102–118. Springer, 2016. 6

    2016

  30. [38]

    Source- free domain adaptation for rgb-d semantic segmentation with vision transformers

    Giulia Rizzoli, Donald Shenaj, and Pietro Zanuttigh. Source- free domain adaptation for rgb-d semantic segmentation with vision transformers. In Proceedings of the IEEE/CVF Win- ter Conference on Applications of Computer Vision (WACV) Workshops, pages 615–624, 2024. 3

    2024

  31. [39]

    Source- free domain adaptation for rgb-d semantic segmentation with vision transformers

    Giulia Rizzoli, Donald Shenaj, and Pietro Zanuttigh. Source- free domain adaptation for rgb-d semantic segmentation with vision transformers. In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, pages 615– 624, 2024. 3

    2024

  32. [40]

    Acdc: The adverse conditions dataset with correspondences for se- mantic driving scene understanding

    Christos Sakaridis, Dengxin Dai, and Luc Van Gool. Acdc: The adverse conditions dataset with correspondences for se- mantic driving scene understanding. In ICCV, pages 10745– 10755, 2021. 6

    2021

  33. [41]

    Open domain generalization with domain-augmented meta-learning

    Yang Shu, Zhangjie Cao, Chenyu Wang, Jianmin Wang, and Mingsheng Long. Open domain generalization with domain-augmented meta-learning. In CVPR, pages 9619– 9628, 2021. 2

    2021

  34. [42]

    Crossnorm and selfnorm for generaliza- tion under distribution shifts

    Zhiqiang Tang, Yunhe Gao, Yi Zhu, Zhi Zhang, Mu Li, and Dimitris Metaxas. Crossnorm and selfnorm for generaliza- tion under distribution shifts. In ICCV, 2021. 1

    2021

  35. [43]

    Cgfnet: Cross-guided fusion network for rgb- t salient object detection

    Jie Wang, Kechen Song, Yanqi Bao, Liming Huang, and Yunhui Yan. Cgfnet: Cross-guided fusion network for rgb- t salient object detection. IEEE TCSVT, 32(5):2949–2961,

  36. [44]

    Stronger fewer & superior: Harnessing vision foundation models for domain generalized semantic segmentation

    Zhixiang Wei, Lin Chen, Yi Jin, Xiaoxiao Ma, Tianle Liu, Pengyang Ling, Ben Wang, Huaian Chen, and Jinjin Zheng. Stronger fewer & superior: Harnessing vision foundation models for domain generalized semantic segmentation. In CVPR, pages 28619–28630, 2024. 1, 2, 4, 6, 7, 8

    2024

  37. [45]

    Sepico: Semantic-guided pixel contrast for domain adaptive semantic segmentation

    Binhui Xie, Shuang Li, Mingjia Li, Chi Harold Liu, Gao Huang, and Guoren Wang. Sepico: Semantic-guided pixel contrast for domain adaptive semantic segmentation. IEEE TPAMI, 45(7):9004–9021, 2023. 1

    2023

  38. [46]

    Dirl: Domain-invariant representation learning for generalizable semantic segmentation

    Qi Xu, Liang Yao, Zhengkai Jiang, Guannan Jiang, Wen- qing Chu, Wenhui Han, Wei Zhang, Chengjie Wang, and Ying Tai. Dirl: Domain-invariant representation learning for generalizable semantic segmentation. In AAAI, pages 2884– 2892, 2022. 6

    2022

  39. [47]

    Dirl: Domain-invariant representation learning for gen- eralizable semantic segmentation

    Qi Xu, Liang Yao, Zhengkai Jiang, Guannan Jiang, Wenqing Chu, Wenhui Han, Wei Zhang, Chengjie Wang, and Ying Tai. Dirl: Domain-invariant representation learning for gen- eralizable semantic segmentation. InAAAI, page 2884–2892,

  40. [48]

    Pixel difference convolutional network for rgb-d semantic segmentation

    Jun Yang, Lizhi Bai, Yaoru Sun, Chunqi Tian, Maoyu Mao, and Guorun Wang. Pixel difference convolutional network for rgb-d semantic segmentation. IEEE TCSVT, 34(3):1481– 1492, 2024. 3

    2024

  41. [49]

    Depth anything: Unleashing the power of large-scale unlabeled data

    Lihe Yang, Bingyi Kang, Zilong Huang, Xiaogang Xu, Jiashi Feng, and Hengshuang Zhao. Depth anything: Unleashing the power of large-scale unlabeled data. In CVPR, 2024. 2

    2024

  42. [50]

    Depth any- thing v2

    Lihe Yang, Bingyi Kang, Zilong Huang, Zhen Zhao, Xiao- gang Xu, Jiashi Feng, and Hengshuang Zhao. Depth any- thing v2. NeurIPS, 2024. 2

    2024

  43. [51]

    Bdd100k: A diverse driving dataset for heterogeneous multitask learning

    Fisher Yu, Haofeng Chen, Xin Wang, Wenqi Xian, Yingying Chen, Fangchen Liu, Vashisht Madhavan, and Trevor Dar- rell. Bdd100k: A diverse driving dataset for heterogeneous multitask learning. In CVPR, pages 2636–2645, 2020. 6

    2020

  44. [52]

    Domain randomization and pyramid consistency: Simulation-to-real generalization without accessing target domain data

    Xiangyu Yue, Yang Zhang, Sicheng Zhao, Alberto Sangiovanni-Vincentelli, Kurt Keutzer, and Boqing Gong. Domain randomization and pyramid consistency: Simulation-to-real generalization without accessing target domain data. In ICCV, pages 2100–2110, 2019. 4, 6 10

    2019

  45. [53]

    Prototypical cross-domain self-supervised learn- ing for few-shot unsupervised domain adaptation

    Xiangyu Yue, Zangwei Zheng, Shanghang Zhang, Yang Gao, Trevor Darrell, Kurt Keutzer, and Alberto Sangiovanni Vincentelli. Prototypical cross-domain self-supervised learn- ing for few-shot unsupervised domain adaptation. In CVPR, pages 13834–13844, 2021. 1, 2

    2021

  46. [54]

    Cmx: Cross-modal fusion for rgb-x semantic segmentation with transformers

    Jiaming Zhang, Huayao Liu, Kailun Yang, Xinxin Hu, Ruip- ing Liu, and Rainer Stiefelhagen. Cmx: Cross-modal fusion for rgb-x semantic segmentation with transformers. IEEE Transactions on Intelligent Transportation Systems, 24(12): 14679–14694, 2023. 3

    2023

  47. [55]

    Unseen object instance segmentation with fully test-time rgb-d embeddings adaptation

    Lu Zhang, Siqi Zhang, Xu Yang, Hong Qiao, and Zhiy- ong Liu. Unseen object instance segmentation with fully test-time rgb-d embeddings adaptation. In 2023 IEEE In- ternational Conference on Robotics and Automation (ICRA), pages 4945–4952, 2023. 3

    2023

  48. [56]

    Towards unsupervised do- main generalization

    Xingxuan Zhang, Linjun Zhou, Renzhe Xu, Peng Cui, Zheyan Shen, and Haoxin Liu. Towards unsupervised do- main generalization. In CVPR, pages 4900–4910, 2022. 2

    2022

  49. [57]

    Style-hallucinated dual consistency learning for domain generalized semantic segmentation

    Yuyang Zhao, Zhun Zhong, Na Zhao, Nicu Sebe, and Gim Hee Lee. Style-hallucinated dual consistency learning for domain generalized semantic segmentation. In ECCV, pages 535–552. Springer, 2022. 4, 6

    2022

  50. [58]

    Adversarial style augmentation for domain general- ized urban-scene segmentation

    Zhun Zhong, Yuyang Zhao, Gim Hee Lee, and Nicu Sebe. Adversarial style augmentation for domain general- ized urban-scene segmentation. NeurIPS, 35:338–350, 2022. 6, 8

    2022

  51. [59]

    Ccafnet: Crossflow and cross-scale adaptive fusion network for detecting salient objects in rgb-d images.IEEE TMM, 24: 2192–2204, 2022

    Wujie Zhou, Yun Zhu, Jingsheng Lei, Jian Wan, and Lu Yu. Ccafnet: Crossflow and cross-scale adaptive fusion network for detecting salient objects in rgb-d images.IEEE TMM, 24: 2192–2204, 2022. 3

    2022

  52. [60]

    Cmpffnet: Cross-modal and progressive feature fusion net- work for rgb-d indoor scene semantic segmentation

    Wujie Zhou, Yuxiang Xiao, Weiqing Yan, and Lu Yu. Cmpffnet: Cross-modal and progressive feature fusion net- work for rgb-d indoor scene semantic segmentation. IEEE Transactions on Automation Science and Engineering , 21 (4):5523–5533, 2024. 3 11

    2024

Pith tools

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