Pith. sign in

REVIEW 3 major objections 5 minor 45 references

BEVCon: Advancing Bird's Eye View Perception with Contrastive Learning

T0 review · 3 major / 5 minor · reviewed 2026-08-05 · deepseek-v4-flash

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

Pith's one-line read Adding two contrastive losses—one on BEV instance features, one on perspective image regions—improves 3D detection by up to 2.4 mAP on nuScenes without extra data or labels.

desk verdict Useful training-time contrastive recipe for BEV detectors, but the headline numbers don't match the paper's own repeated-run means, and the ablation story is shakier than the text claims. read the letter →

arxiv 2508.04702 v1 pith:UPHQGTBU submitted 2025-08-06 cs.CV

Ziyang Leng , Jiawei Yang , Zhicheng Ren , Bolei Zhou This is my paper
classification cs.CV
keywords bird'seyeviewperceptioncontrastivelearning3DobjectdetectionautonomousdrivingrepresentationnuScenesmulti-camera
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

BEVCon claims that the bottleneck in camera-based Bird's Eye View perception is not just the encoder architecture or the detection head, but the quality of the learned feature representations themselves. The paper inserts two contrastive learning modules into existing BEV detectors: one pulls together BEV-pooled features of the same object from two augmented views while pushing apart different objects, and the other does the same for region-pooled features in the perspective camera images. Trained jointly with the detection loss on nuScenes, this raises mAP by up to 2.4 points (BEVFormer-tiny from 0.252 to 0.276) and improves accuracy across BEVDet, BEVDet4D, BEVFormer, and Sparse4D—families with different view transform mechanisms. The reason to care is that the gains come without new data, new labels, or architectural changes, suggesting representation-level supervision is a widely applicable complement to task-specific design. The paper also reports that image-level contrastive pre-training on driving data fails to help, which motivates the object-level, dense design.

What carries the argument

The load-bearing machinery is the pair of contrastive losses operating at object level. The instance feature contrast loss $\mathcal{L}_{\text{in}}$ uses RoI Align to pool features from BEV grids $B$ and $B'$ at the same ground-truth locations, forming positive pairs per instance and negatives across instances, with a SimCLR-style InfoNCE loss and cosine similarity; it is applied at multiple BEV encoder layers with an exponential scale $1/\epsilon^{N_{\text{layer}}-l}$. The perspective regional contrast loss $\mathcal{L}_{\text{pers}}$ pools multi-level image features with scale-aware pooling (shrinking boxes by $\gamma=0.6$) and contrasts them across the two augmented views. An EMA-updated

What would settle it

Project every 3D ground-truth box center through the view transform into both augmented BEV grids $B$ and $B'$ and compute the fraction of instance pairs whose RoI-Align pooling regions cover the same grid cell; if that fraction is much less than 1, the alignment assumption fails and the reported gains cannot be attributed to the contrastive mechanism as described. Alternatively, train with independent random BEV rotations for the two views so positives are deliberately misaligned: a significant mAP drop relative to the aligned setting would confirm that alignment, not the contrastive objectiv

Watch

Extended reading notes

Core claim

On the paper's own terms, the central discovery is that dense, annotation-guided contrastive learning—not generic image-level contrast—benefits BEV perception. The instance feature contrast module takes BEV features $B$ and $B'$ produced from two augmented views of the same frame, pools per-object features at ground-truth locations with RoI Align, and maximizes cosine similarity between matching instances while treating other instances as negatives. The perspective regional contrast module applies the same idea to multi-scale image features, using scale-aware pooling ($\gamma=0.6$) to avoid fusing overlapping bounding boxes. Both losses are added to the detection loss with EMA-updated backbo

Load-bearing premise

The load-bearing premise is that the data-augmentation pipeline keeps each object in exactly the same BEV grid cell across the two augmented views; if cropping, rotation, or BEV-space augmentation breaks this correspondence, the positive pairs used by the instance contrast loss would be mismatched and the added loss would actively harm learning.

Editorial extensions

If this is right

  • If BEVCon's claim holds, any existing camera-based BEV detector can be improved without new data or labels by adding these two losses, as demonstrated on BEVDet, BEVDet4D, BEVFormer, and Sparse4D.
  • The consistent NDS and mAP gains across depth-based, dense-query, and sparse-query view transforms suggest the improvement is in shared representation quality rather than in any single architecture-specific component.
  • The reduction in mATE, mAOE, and mAVE errors implies contrastive supervision sharpens features relevant for localization, orientation, and velocity—properties that should transfer to downstream tracking and planning.
  • The failure of image-level contrastive pre-training/joint training indicates that object-level, annotation-anchored contrast is the operative ingredient, so future contrastive designs for driving perception should focus at instance/region level.

Reading between the lines

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

  • Editorial inference: the same instance-level contrastive losses could be extended to other BEV tasks—semantic segmentation, trajectory prediction, and planning—since the representation-level supervision is not tied to box detection; a natural test is attaching the modules to BEV segmentation or planning models.
  • The alignment assumption implies a portability constraint: for a BEV method whose view transform is not a fixed mapping (e.g., recurrent or sampling-based queries), the two augmented views may not yield identical BEV grid positions, so the positive-pair construction would need re-derivation rather than direct copying.
  • A cheap diagnostic extension would be to measure positive-pair feature similarity during training: if the contrastive loss is working as intended, same-instance pairs across augmented views should become substantially more similar than cross-instance pairs, and this gap should correlate with the mAP gain.
  • The gains reported on nuScenes may understate or overstate the value on datasets with different object density, resolution, or camera overlap; the framework's reliance on ground-truth boxes for pooling means label noise or missing annotations at train time could directly degrade the contrast signal.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 5 minor

Summary. The paper introduces BEVCon, a training-time contrastive learning framework for BEV perception. Two auxiliary losses are added to a standard detection loss: an instance feature contrast loss computed on RoI-pooled BEV features from two augmented views, and a perspective regional contrast loss computed on multi-level image features. The method is evaluated on nuScenes 3D detection with BEVDet/BEVDet4D, BEVFormer-tiny/base, and Sparse4D, reporting consistent gains and an ablation study with repeated runs. The central claim is that these objectives, without extra data or labels, improve both the BEV encoder and the image backbone, giving up to +2.4 mAP over BEVFormer-tiny.

Significance. If the reported gains are reproducible, BEVCon would be a useful, architecture-agnostic training objective for multi-camera BEV detectors. The paper's strengths are its clear motivation, the breadth of detector families tested, and the fact that the main ablations are run multiple times with reported standard deviations. The code and models are promised open-source, which is valuable for verification. However, the central quantitative claim is weakened by an internal inconsistency between the headline tables and the controlled repeated-run tables, and the augmentation-alignment assumption underlying the contrastive pairs is underspecified. With those issues resolved, the contribution could be significant for the RA-L community.

major comments (3)
  1. [§IV.D, Table III] The headline result is not internally reproducible. Table I reports BEVFormer-tiny+BEVCon as 0.375 NDS / 0.276 mAP, and Table II reports BEVFormer-base+BEVCon as 0.528 NDS / 0.424 mAP. However, the controlled ablation in Table III reports 'Contrast Framework' for ResNet-50 as 0.3709±0.0002 NDS / 0.2735±0.0002 mAP, and the full configuration in Table IV(a) (all five components) as 0.371 / 0.274. The differences (0.004 NDS, 0.002 mAP) are 10–20 times the reported standard errors, so they cannot be run-to-run noise. The same pattern holds for ResNet-101-DCN: Table II gives 0.528/0.424 vs 0.525±0.0004/0.422±0.0005 in Tables III/IV(b). The abstract's 'up to +2.4% mAP' is calibrated to the Table I value; the repeated-run numbers would support at most +2.2% mAP, and the NDS gain would be 1.7 points rather than 2.1. Please state which numbers are the official results, whether Tables I/II are bes
  2. [§IV.D, Table III] The ablation results do not support the sentence 'the combination of them could further improve the performance.' For ResNet-50, Instance Contrast alone gives NDS 0.375, Perspective Contrast alone gives 0.373, but Contrast Framework gives 0.371—lower than either individual module on NDS, despite a tiny mAP improvement (0.272→0.273→0.274). If NDS is the primary metric, the two modules are not complementary in this configuration. The same issue appears in Table IV(a), where the full model (0.371/0.274) does not dominate the 'Ins+Align+MLC' row (0.375/0.272) on NDS. Please report per-component significance and clarify whether the claimed synergy refers only to mAP.
  3. [§III.C-D, §IV.B] The correctness of the method depends on exact BEV-space correspondence between the two augmented views: Eq. (1) treats fi and f'i as features of the same grid cell/instance. Section IV.B only asserts that 'the whole augmentation process ensures that each instance feature pair ... is located in the same position in BEV space', but does not specify how the 2D image augmentations (random resize, crop, rotation, distortion) are coupled with the BEV augmentations, nor how the 3D annotations are transformed for RoI pooling. A subtle misalignment—e.g., a per-camera resize that changes the projection—would make positive pairs mismatched and could actively harm training. Please provide the exact augmentation protocol (including alignment of 3D boxes with the BEV grid), and an empirical check such as the localization error of pooled features or a sensitivity study to small perturbations.
minor comments (5)
  1. [Tables I, II, III] Baseline rows are reported without error bars, while the proposed method's repeated-run standard deviations are at the 10^-4 level. Please report the number of seeds and mean±std for the baselines as well, since the reported gains are small in several configurations (e.g., BEVDet4D +0.3 mAP).
  2. [Table IV] The checkmark formatting is difficult to parse; each row's component combination should be explicit (e.g., separate columns for Ins, Align, MLC, Pers, Scale with clear ✓/✗ entries). As printed, several rows are ambiguous.
  3. [§IV.B] The sentence 'τ is set respectively to 0.2 and 0.05 for the two image backbones' should specify which backbone uses which value, and whether the same temperature is used for both contrast modules.
  4. [§III-C, Eq. (2)] Define Nlayer explicitly and clarify whether the exponential scaling is applied to all transformer layers or only intermediate ones. No ablation or sensitivity study is provided for the hyperparameter ε.
  5. [Table V] The 'ACO' dataset is not defined in the main text. Please give a reference and describe the pretraining setup (epochs, optimizer, views) so the comparison is reproducible.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: BEVCon's contrastive losses are independent objectives and the mAP/NDS gains are measured by an external benchmark; internal number mismatches are a reproducibility concern, not a derivation collapse.

full rationale

BEVCon proposes two auxiliary contrastive losses (instance feature contrast in Eq. 1 and perspective regional contrast) added to the detection loss in Eq. 3. The claimed contribution is that jointly optimizing these losses improves 3D detection. The mAP/NDS numbers are evaluated on nuScenes val, an external benchmark, against the original detection losses. There is no step in the derivation where the output (detection performance) is defined in terms of the input (contrastive loss), nor is any fitted parameter renamed as a prediction; the contrastive losses are standard objective functions and the reported metrics are independent measurements. The cited related work, including self-citations [1], [45], is background and not load-bearing: no central premise is justified solely by a self-citation. The internal inconsistency between headline numbers in Tables I/II and repeated-run means in Tables III/IV (e.g., BEVFormer-tiny mAP 0.276 in Table I vs 0.2735±0.0002 in Table III) is a reproducibility/correctness concern, not a circularity. Augmentation alignment is an assumed implementation property, not an equation that reduces the result to itself. Therefore no circular step can be exhibited.

Assumptions & free parameters 5 free parameters · 4 assumptions · 0 invented entities

The central claim rests on standard contrastive learning math, on the accuracy of nuScenes annotations, and on a specific augmentation-alignment engineering assumption. The free parameters are all hyperparameters, not physical constants. There are no invented entities.

free parameters (5)
  • Contrastive temperature tau = 0.2 (ResNet-50), 0.05 (ResNet-101-DCN)
    Used in the InfoNCE-style losses in Eq. (1); values differ per backbone with no sensitivity analysis, likely tuned on nuScenes val.
  • Loss weights lambda_in, lambda_pers = 1.0, 1.0
    Weights in joint loss Eq. (3); set to 1 without reported tuning.
  • Center scale factor gamma = 0.6
    Downscales 2D bounding boxes in scale-aware pooling to avoid overlapping objects; chosen by hand.
  • Multi-layer loss scale epsilon = 0.5
    Controls exponential weighting of instance contrast loss across transformer layers in Eq. (2); no sensitivity study.
  • EMA momentum = not reported
    The EMA-updated image backbone in Section III-B requires a momentum hyperparameter, but no value is given; replication depends on it.
assumptions (4)
  • domain assumption The view transform mapping M is shared between the two augmented views
    Section III-B states a shared view transform module produces B and B', assuming the same M works for both augmented inputs; if M is not symmetric under augmentation, contrastive pairs are biased.
  • domain assumption Ground-truth 3D and 2D annotations are accurate enough to define positive/negative instance pairs
    The instance and perspective contrast modules use annotations A and A2D (Sections III-C, III-D); annotation noise or missing labels would create false pairs.
  • domain assumption The augmentation pipeline preserves BEV location alignment for every instance
    Section IV.B states that augmentations ensure each instance pair is at the same position in BEV space; this is necessary for RoI Align pooling to extract corresponding features.
  • ad hoc to paper Exponential scaling epsilon^(Nlayer-l) is a valid weighting of layer-wise losses
    Eq. (2) introduces epsilon with no derivation; it is an ad hoc design choice.

how reviews work

0 comments
Cite this review

Pith. "Pith review of BEVCon: Advancing Bird's Eye View Perception with Contrastive Learning." pith.science (2026). https://pith.science/paper/UPHQGTBU

@misc{pith2026250804702,
  author       = {Pith},
  title        = {Pith review of: BEVCon: Advancing Bird's Eye View Perception with Contrastive Learning},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/UPHQGTBU}},
  note         = {Machine review of arXiv:2508.04702}
}
read the original abstract

We present BEVCon, a simple yet effective contrastive learning framework designed to improve Bird's Eye View (BEV) perception in autonomous driving. BEV perception offers a top-down-view representation of the surrounding environment, making it crucial for 3D object detection, segmentation, and trajectory prediction tasks. While prior work has primarily focused on enhancing BEV encoders and task-specific heads, we address the underexplored potential of representation learning in BEV models. BEVCon introduces two contrastive learning modules: an instance feature contrast module for refining BEV features and a perspective view contrast module that enhances the image backbone. The dense contrastive learning designed on top of detection losses leads to improved feature representations across both the BEV encoder and the backbone. Extensive experiments on the nuScenes dataset demonstrate that BEVCon achieves consistent performance gains, achieving up to +2.4% mAP improvement over state-of-the-art baselines. Our results highlight the critical role of representation learning in BEV perception and offer a complementary avenue to conventional task-specific optimizations.

Figures

Figures reproduced from arXiv: 2508.04702 by the authors.

Figure 1
Figure 1. The overall framework of BEVCon, which consists [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Illustration of two contrastive learning modules. Red [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. A qualitative comparison of the detection results [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

45 extracted references · 8 linked inside Pith

  1. [1]

    Cross-view semantic segmentation for sensing surroundings,

    B. Pan, J. Sun, H. Y . T. Leung, A. Andonian, and B. Zhou, “Cross-view semantic segmentation for sensing surroundings,” IEEE Robotics and Automation Letters, vol. 5, no. 3, pp. 4867–4873, 2020

    2020

  2. [2]

    Lift, splat, shoot: Encoding images from arbitrary camera rigs by implicitly unprojecting to 3d,

    J. Philion and S. Fidler, “Lift, splat, shoot: Encoding images from arbitrary camera rigs by implicitly unprojecting to 3d,” in European Conference on Computer Vision (ECCV) , 2020, p. 194–210

    2020

  3. [3]

    Bevformer v2: Adapting modern image backbones to bird’s-eye-view recognition via perspective supervision,

    C. Yang et al. , “Bevformer v2: Adapting modern image backbones to bird’s-eye-view recognition via perspective supervision,” 2022

    2022

  4. [4]

    Covis-net: A co- operative visual spatial foundation model for multi-robot applications,

    J. Blumenkamp, S. Morad, J. Gielis, and A. Prorok, “Covis-net: A co- operative visual spatial foundation model for multi-robot applications,” arXiv preprint arXiv:2405.01107 , 2024

    arXiv 2024

  5. [5]

    Bevdet: High-performance multi-camera 3d object detection in bird-eye-view,

    J. Huang et al. , “Bevdet: High-performance multi-camera 3d object detection in bird-eye-view,” arXiv preprint arXiv:2112.11790 , 2021

    arXiv 2021

  6. [6]

    Bevfusion: Multi-task multi-sensor fusion with unified bird’s-eye view representation,

    Z. Liu et al. , “Bevfusion: Multi-task multi-sensor fusion with unified bird’s-eye view representation,” in 2023 IEEE International Conference on Robotics and Automation (ICRA) , 2023, pp. 2774–2781

    2023

  7. [7]

    Simple-bev: What really matters for multi-sensor bev perception?

    A. W. Harley et al., “Simple-bev: What really matters for multi-sensor bev perception?” in 2023 IEEE International Conference on Robotics and Automation (ICRA) , 2023, pp. 2759–2765

    2023

  8. [8]

    Bevdepth: Acquisition of reliable depth for multi-view 3d object detection,

    Y . Li et al. , “Bevdepth: Acquisition of reliable depth for multi-view 3d object detection,” Proceedings of the AAAI Conference on Artificial Intelligence, vol. 37, no. 2, pp. 1477–1485, 2023

    2023

Show all 45 references
  1. [9]

    Petr: Position embedding trans- formation for multi-view 3d object detection,

    Y . Liu, T. Wang, X. Zhang, and J. Sun, “Petr: Position embedding trans- formation for multi-view 3d object detection,” in European Conference on Computer Vision . Springer, 2022, pp. 531–548

    2022

  2. [10]

    A simple framework for contrastive learning of visual representations,

    T. Chen, S. Kornblith, M. Norouzi, and G. Hinton, “A simple framework for contrastive learning of visual representations,” in International conference on machine learning . PMLR, 2020, pp. 1597–1607

    2020

  3. [11]

    Momentum contrast for unsupervised visual representation learning,

    K. He et al., “Momentum contrast for unsupervised visual representation learning,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , 2020, pp. 9729–9738

    2020

  4. [12]

    Supervised contrastive learning,

    P. Khosla et al., “Supervised contrastive learning,” Advances in neural information processing systems , vol. 33, pp. 18 661–18 673, 2020

    2020

  5. [13]

    Contrastive learning for object detection,

    R. Balasubramanian and K. Rathore, “Contrastive learning for object detection,” arXiv preprint arXiv:2208.06412 , 2022

    2022 arXiv

  6. [14]

    Icpc: Instance-conditioned prompting with contrastive learning for semantic segmentation,

    C. Yu, Q. Zhou, Z. Wang, and F. Wang, “Icpc: Instance-conditioned prompting with contrastive learning for semantic segmentation,” arXiv preprint arXiv:2308.07078, 2023

    2023 arXiv

  7. [15]

    nuscenes: A multimodal dataset for autonomous driving,

    H. Caesar et al. , “nuscenes: A multimodal dataset for autonomous driving,” in CVPR, 2020

    2020

  8. [16]

    Scalability in perception for autonomous driving: Waymo open dataset,

    P. Sun et al., “Scalability in perception for autonomous driving: Waymo open dataset,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , June 2020

    2020

  9. [17]

    Argoverse 2: Next generation datasets for self-driving perception and forecasting,

    B. Wilson et al., “Argoverse 2: Next generation datasets for self-driving perception and forecasting,” in Proceedings of the Neural Information Processing Systems Track on Datasets and Benchmarks , vol. 1, 2021

    2021

  10. [18]

    Bridgedata v2: A dataset for robot learning at scale,

    H. R. Walke et al., “Bridgedata v2: A dataset for robot learning at scale,” in Conference on Robot Learning . PMLR, 2023, pp. 1723–1736

    2023

  11. [19]

    Rh20t: A comprehensive robotic dataset for learning diverse skills in one-shot,

    H.-S. Fang et al., “Rh20t: A comprehensive robotic dataset for learning diverse skills in one-shot,” in 2024 IEEE International Conference on Robotics and Automation (ICRA) . IEEE, 2024, pp. 653–660

    2024

  12. [20]

    ImageNet: A Large-Scale Hierarchical Image Database,

    J. Deng, W. Dong, R. Socher, L.-J. Li, K. Li, and L. Fei-Fei, “ImageNet: A Large-Scale Hierarchical Image Database,” in CVPR09, 2009

    2009

  13. [21]

    Bevformer: Learning bird’s-eye-view representation from multi-camera images via spatiotemporal transformers,

    Z. Li et al., “Bevformer: Learning bird’s-eye-view representation from multi-camera images via spatiotemporal transformers,” in European Conference on Computer Vision (ECCV) , 2022, p. 1–18

    2022

  14. [22]

    Sparse4d: Multi-view 3d object detection with sparse spatial-temporal fusion,

    X. Lin et al. , “Sparse4d: Multi-view 3d object detection with sparse spatial-temporal fusion,” arXiv preprint arXiv:2211.10581 , 2022

    2022 arXiv

  15. [23]

    Detr3d: 3d object detection from multi-view images via 3d-to-2d queries,

    Y . Wang et al. , “Detr3d: 3d object detection from multi-view images via 3d-to-2d queries,” in Conference on Robot Learning. PMLR, 2022, pp. 180–191

    2022

  16. [24]

    Understanding bird’s-eye view of road semantics using an onboard camera,

    Y . B. Can et al., “Understanding bird’s-eye view of road semantics using an onboard camera,” IEEE Robotics and Automation Letters , vol. 7, no. 2, pp. 3302–3309, 2022

    2022

  17. [25]

    Petrv2: A unified framework for 3d perception from multi-camera images,

    Y . Liu et al. , “Petrv2: A unified framework for 3d perception from multi-camera images,” in Proceedings of the IEEE/CVF International Conference on Computer Vision , 2023, pp. 3262–3272

    2023

  18. [26]

    Exploring object-centric temporal modeling for efficient multi-view 3d object detection,

    S. Wang et al., “Exploring object-centric temporal modeling for efficient multi-view 3d object detection,” arXiv preprint arXiv:2303.11926, 2023

    2023 arXiv

  19. [27]

    Polarformer: Multi-camera 3d object detection with polar transformer,

    Y . Jiang et al. , “Polarformer: Multi-camera 3d object detection with polar transformer,” in Proceedings of the AAAI Conference on Artificial Intelligence, 2023

    2023

  20. [28]

    Tig-bev: Multi-view bev 3d object detection via target inner-geometry learning,

    P. Huang et al., “Tig-bev: Multi-view bev 3d object detection via target inner-geometry learning,” 2022

    2022

  21. [29]

    Sa-bev: Generating semantic- aware bird’s-eye-view feature for multi-view 3d object detection,

    J. Zhang, Y . Zhang, Q. Liu, and Y . Wang, “Sa-bev: Generating semantic- aware bird’s-eye-view feature for multi-view 3d object detection,” 2023

    2023

  22. [30]

    Fb- bev: Bev representation from forward-backward view transformations,

    Z. Li, Z. Yu, W. Wang, A. Anandkumar, T. Lu, and J. M. ´Alvarez, “Fb- bev: Bev representation from forward-backward view transformations,” IEEE/CVF International Conference on Computer Vision (ICCV) , 2023

    2023

  23. [31]

    Time will tell: New outlooks and a baseline for temporal multi-view 3d object detection,

    J. Park et al., “Time will tell: New outlooks and a baseline for temporal multi-view 3d object detection,” International Conference on Learning Representations, 2023

    2023

  24. [32]

    Swin transformer: Hierarchical vision transformer using shifted windows,

    Z. Liu et al., “Swin transformer: Hierarchical vision transformer using shifted windows,” in Proceedings of the IEEE/CVF international con- ference on computer vision , 2021, pp. 10 012–10 022

    2021

  25. [33]

    Internimage: Exploring large-scale vision foundation models with deformable convolutions,

    W. Wang et al. , “Internimage: Exploring large-scale vision foundation models with deformable convolutions,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , 2023, pp. 14 408–14 419

    2023

  26. [34]

    Bootstrap your own latent-a new approach to self- supervised learning,

    J.-B. Grill et al. , “Bootstrap your own latent-a new approach to self- supervised learning,” Advances in neural information processing sys- tems, vol. 33, pp. 21 271–21 284, 2020

    2020

  27. [35]

    Barlow twins: Self-supervised learning via redundancy reduction,

    J. Zbontar, L. Jing, I. Misra, Y . LeCun, and S. Deny, “Barlow twins: Self-supervised learning via redundancy reduction,” in International Conference on Machine Learning . PMLR, 2021, pp. 12 310–12 320

    2021

  28. [36]

    Semi-supervised semantic segmentation with pixel- level contrastive learning from a class-wise memory bank,

    I. Alonso et al. , “Semi-supervised semantic segmentation with pixel- level contrastive learning from a class-wise memory bank,” in Proceed- ings of the IEEE/CVF International Conference on Computer Vision , 2021, pp. 8219–8228

    2021

  29. [37]

    Pixel contrastive-consistent semi-supervised semantic segmentation,

    Y . Zhong et al., “Pixel contrastive-consistent semi-supervised semantic segmentation,” in Proceedings of the IEEE/CVF International Confer- ence on Computer Vision , 2021, pp. 7273–7282

    2021

  30. [38]

    Coin: Contrastive instance feature mining for outdoor 3d object detection with very limited annotations,

    Q. Xia et al. , “Coin: Contrastive instance feature mining for outdoor 3d object detection with very limited annotations,” in Proceedings of the IEEE/CVF International Conference on Computer Vision , 2023, pp. 6254–6263

    2023

  31. [39]

    Planning-oriented autonomous driving,

    Y . Hu et al. , “Planning-oriented autonomous driving,” in 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). IEEE, 2023

    2023

  32. [40]

    End-to-end object detection with transformers,

    N. Carion et al., “End-to-end object detection with transformers,” 2020

    2020

  33. [41]

    Mask r-cnn,

    K. He et al. , “Mask r-cnn,” in Proceedings of the IEEE international conference on computer vision , 2017, pp. 2961–2969

    2017

  34. [42]

    Bevdet4d: Exploit temporal cues in multi- camera 3d object detection,

    J. Huang and G. Huang, “Bevdet4d: Exploit temporal cues in multi- camera 3d object detection,” arXiv preprint arXiv:2203.17054 , 2022

    2022 arXiv

  35. [43]

    Deep residual learning for image recognition,

    K. He, X. Zhang, S. Ren, and J. Sun, “Deep residual learning for image recognition,” in Proceedings of the IEEE conference on computer vision and pattern recognition , 2016, pp. 770–778

    2016

  36. [44]

    Improved baselines with mo- mentum contrastive learning,

    X. Chen, H. Fan, R. Girshick, and K. He, “Improved baselines with mo- mentum contrastive learning,” arXiv preprint arXiv:2003.04297 , 2020

    2003 arXiv

  37. [45]

    Learning to drive by watching youtube videos: Action-conditioned contrastive policy pretraining,

    Q. Zhang, Z. Peng, and B. Zhou, “Learning to drive by watching youtube videos: Action-conditioned contrastive policy pretraining,” European Conference on Computer Vision (ECCV) , 2022

    2022

Pith tools

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