Pith. sign in

REVIEW 3 major objections 5 minor 43 references

DySS: Dynamic Queries and State-Space Learning for Efficient 3D Object Detection from Multi-Camera Videos

T0 review · 3 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read The paper argues that DySS makes camera-only 3D detection from multi-camera video both more accurate and faster by summarizing scene history in a state-space model and by dynamically pruning, merging, and splitting the detection queries…

desk verdict Solid nuScenes SOTA claim with a reproducible gap: the dynamic query training signal is never specified, so the speed-up story is unattributable as written. read the letter →

arxiv 2506.10242 v1 pith:4PYWXD5N submitted 2025-06-11 cs.CV

Rajeev Yasarla , Shizhong Han , Hong Cai , Fatih Porikli This is my paper
classification cs.CV
keywords dynamicqueriesstate-spacemodel3Dobjectdetectionmulti-cameravideoBEVperceptionnuScenestemporalmodelingquerypruning
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

DySS sets out to show that camera-only 3D detection from multi-camera video does not need a fixed, large set of queries or a dense BEV grid to be state of the art. Instead, a state-space model reads the image features sampled by the queries over consecutive frames and compresses them into a hidden state that summarizes motion and scene correspondence; auxiliary losses on masked reconstruction and future-frame prediction train this state to be informative. The decoder then edits its queries between layers, merging similar ones, removing redundant ones, and splitting others, shrinking the query set from 900 to 269 on the validation split and from 1600 to 729 on the test split. If the claim holds, the practical consequence is that autonomous-driving perception can get higher accuracy at real-time speed: 56.2 NDS and 46.2 mAP at 33 FPS on the validation split, and 65.31 NDS and 57.4 mAP on the test benchmark.

What carries the argument

The load-bearing machinery is the recurrent state update $h(t) = \boldsymbol A h(t-1) + \boldsymbol B x(t)$, $y(t) = \boldsymbol C h(t)$, realized with Mamba-style selective state-space blocks; one block operates in the time domain and one in the frequency domain via FFT. The SSM consumes sampled features $F_{t=i}$ plus predicted features $\tilde{F}_{t=i+1}$ and is trained with reconstruction and prediction losses $L_r$ and $L_f$ in Eq. (2), which is what makes the hidden state encode motion rather than simply pass features through. The second mechanism is the dynamic query editor: after cross-attention with state-space features, linear layers read the query features and a covariance matrix $\boldsymbol C_q$ to emit merge, remove, and split labels, removing 20-30% of queries and splitting at most 5%, so the query count falls adaptively through the six decoder layers.

What would settle it

Run DySS with the learned merge, remove, and split heads replaced by fixed random decisions that produce the same final query count (e.g., always dropping 30% and splitting 5% at random, and merging at random); if NDS and mAP stay within noise of the learned-head model, the reported gains come from the reduced query count or the SSM, not from learned query editing. A complementary check is to inspect the remove labels after training: if queries marked for removal sit as close to ground-truth objects as queries kept, the head is not learning redundancy.

Watch

Extended reading notes

Core claim

The central claim is that temporal context for sparse query detection should be represented as the evolving state of a state-space model rather than as extra attention over many sampled points from many frames. DySS feeds the SSM the features sampled at each time step, together with a predicted version of those features from the previous step, and uses the final hidden state as a compact spatial-temporal summary of the scene. Supervised by masked reconstruction and future-feature prediction during training, the SSM is claimed to capture motion and cross-view correspondence that improves both detection and tracking. On top of this, merge, remove, and split operations driven by query-feature covariances let the decoder maintain a lean set of queries; the paper reports that this dynamic editing keeps accuracy essentially equal to the 900-query baseline while nearly tripling the frame rate, and that the full system outperforms prior state of the art on the nuScenes benchmark.

Load-bearing premise

The load-bearing premise is that the ordinary detection losses alone teach the merge, remove, and split heads to make useful query-editing decisions, because the paper specifies no loss or ground-truth label for those operations.

Editorial extensions

If this is right

  • Temporal history becomes cheap: a compact SSM state replaces the expensive per-query multi-frame sampling and attention of earlier sparse detectors, so longer video context does not multiply compute.
  • Query editing gives an accuracy-preserving speed dial: cutting queries from 900 to 269 raises FPS from 11 to 35 while NDS stays near the full-query baseline.
  • The learned temporal representation transfers to downstream tasks: DySS improves AMOTA over SparseBEV and StreamPETR, not just box accuracy.
  • Camera-only detection reaches a real-time operating point on a single GPU, 33 FPS on an RTX-3080, while setting a new nuScenes test score of 65.31 NDS.

Reading between the lines

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

  • If dynamic query editing is what carries the speed gain, then the merge/remove/split ratios could be made a continuous function of a latency budget, letting one network trade accuracy for speed at inference time.
  • The auxiliary masked-reconstruction and future-prediction losses suggest a self-supervised pretraining recipe: train the SSM on unlabeled driving video with those two losses, then fine-tune with detection losses, which the paper does not explore.
  • A direct test of the query editor would be to supervise merge/remove/split labels with a sparse-set matching loss; lack of such supervision is the main unstated assumption, and adding it would either strengthen or refute the claimed mechanism.
  • Using FFT as one of the two SSM branches hints that spectral features help motion modeling; a natural extension is replacing the fixed transform with a learned frequency filter bank and measuring the change on high-velocity classes.
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. DySS proposes a sparse-query, multi-camera 3D object detector that combines a state-space model (Mamba) with dynamic query operations (merge, remove, split). The state-space model is trained with auxiliary masked-reconstruction and future-prediction losses, and its features are mixed into the query features. The claimed contributions are improved accuracy and efficiency: on nuScenes test the paper reports 65.3 NDS / 57.4 mAP, on val 56.2 NDS / 46.2 mAP at 33 FPS, and an ablation on val shows that reducing queries from 900 to 269 via the dynamic operations roughly preserves accuracy while increasing speed.

Significance. If the claims hold, the paper makes a practical contribution: it shows that a state-space representation can summarize multi-view video efficiently, and that pruning detection queries dynamically can cut compute substantially without losing accuracy. The external nuScenes test-server numbers and the clear ablation table are strengths, and the comparison against SparseBEV and StreamPETR is appropriate. The main reason for caution is that the training signal for the dynamic query module is not specified, so the central efficiency-accuracy trade-off is not yet reproducible from the manuscript.

major comments (3)
  1. [Section 3.1, Eq. (2), Section 4.1] The dynamic query operations are not supervised or otherwise given a training signal. Section 3.1.1-3.1.3 says linear layers produce merge labels, removal labels in [0,1], a removal ratio in [0.2,0.3], split labels in [0,1], and a split count in [0,5], but no loss term or reward is defined for any of these outputs. Section 4.1 lists only the detection losses and the state-space losses Lr and Lf from Eq. (2); none involve merge/remove/split decisions. Because the operations are discrete, detection losses cannot backpropagate through them unless a relaxation or estimator (e.g., Gumbel-Softmax, straight-through, REINFORCE) is described, and none is. This is load-bearing: Table 3 attributes the preserved accuracy and the FPS gain to the dynamic query module, yet the mechanism by which the module learns to merge, remove, and split is absent. The authors should specify the supervision or training procedure for these decisions, or add a control experiment showing that random labels under the same quotas do not produce the same result.
  2. [Section 4.3.2 vs. Table 2 vs. Abstract] The nuScenes test numbers are internally inconsistent. The abstract states 65.31 NDS and 57.4 mAP, Table 2 reports 65.3 NDS and 57.4 mAP, but Section 4.3.2 first says DySS achieves 65.1 NDS and 56.8 mAP and later, in the same paragraph, says it achieves 65.3 NDS and 57.4 mAP. Since the test split result is the paper's headline claim, the correct numbers need to be stated consistently in all three places.
  3. [Section 4.3.1 and Table 3] The comparison against SparseBEV in Section 4.3.1 says DySS is 'more than 20% faster than SparseBEV (27 FPS)', but the SparseBEV baseline in Table 3 is measured at 14 FPS with 900 queries (or 10-11 FPS when the SSM is added). The source of the 27 FPS figure is not given, and the two numbers are not compatible. The speed comparison should state which configuration of SparseBEV is used and under which hardware/software conditions, or the ablation-table baseline should be used as the reference.
minor comments (5)
  1. [Section 3.2] Typo: 'It outputs an enhanced version of the sampled features TF̂_t=1' should presumably be 'TF̂_t=i'.
  2. [References] References [6] and [7] are the same paper (Mamba by Gu and Dao); the duplicate citation should be merged.
  3. [Section 3.1.1] The merge operation is described only as producing a label and an index; the manuscript does not state how the selected queries are actually merged (e.g., weighted averaging of features, which query is kept). This is needed for reproducibility even if the supervision question is resolved.
  4. [Eq. (3)-(4)] The state-space feature is written as S^l_{t=T}, but the decoder layer index l is not defined at that point; it should be introduced or the superscript omitted.
  5. [Table 3] The FPS column shows that adding SSM with identity/FFT drops speed from 14 to 10 FPS, and then dynamic queries raise it to 33-35 FPS; a sentence explaining that the SSM overhead is amortized by the query reduction would help readers interpret the trade-off.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: DySS's performance claims are validated on the external nuScenes benchmark, and its auxiliary losses are training regularizers rather than redefinitions of the detection target.

full rationale

The central claims, including 65.31 NDS and 57.4 mAP on nuScenes test and 56.2 NDS and 46.2 mAP on val at 33 FPS, are evaluated against an external benchmark, so they are not forced by construction. The state-space auxiliary losses in Eq. (2) supervise intermediate feature reconstruction and future prediction against sampled features; these are auxiliary training objectives, not redefinitions of the final detection metric. The dynamic query merge, remove, and split operations are learned modules whose labels are produced by linear layers from query features and covariance matrices; the paper does not specify how these discrete decisions are supervised, which is a reproducibility or underspecification concern rather than circularity, because the final benchmark numbers do not reduce to a fitted input or to an equation identical to the output. The only self-citation, FutureDepth [40], is used as a supporting example that auxiliary supervision helps video depth estimation, and it is not load-bearing for the nuScenes detection result. No circular step can be exhibited from the paper's equations or citations, so the honest finding is no significant circularity.

Assumptions & free parameters 6 free parameters · 4 assumptions · 0 invented entities

The central empirical claim is not a derivation, so the ledger lists the hand-set hyperparameters and unstated empirical assumptions that the reported accuracy and speed depend on. No new physical or conceptual entities are introduced; dynamic query operations are architectural mechanisms, not entities.

free parameters (6)
  • initial query count = 900 (val), 1600 (test)
    Chosen by hand following SparseBEV; performance varies with count as shown in Table 3.
  • state expansion factor = 128
    Set for all experiments without an ablation study.
  • remove ratio bounds = 0.2 to 0.3
    Ad hoc cap on removed queries in Section 3.1.2.
  • split fraction = max 5% of queries
    Ad hoc cap for the split operation in Section 3.1.3.
  • auxiliary loss weights
    No weights for Lr and Lf in Eq. (2) are reported.
  • number of frames = 8
    Default setting with 0.5 second interval; no ablation is shown for this choice.
assumptions (4)
  • domain assumption The state-space model state can summarize spatial-temporal scene information from sampled multi-frame features.
    Invoked in Sections 3.2 and 3.3; supported only by the ablation table, not by independent analysis.
  • ad hoc to paper The covariance matrix of query features captures query similarity well enough to drive merge, remove, and split decisions.
    Introduced in Section 3.1.1 with no justification or analysis of why covariance is the right signal.
  • domain assumption Masked reconstruction and future prediction improve the state-space feature quality enough to help the detection task.
    Self-supervision assumption in Section 3.2; supported by one ablation row comparing with and without the losses.
  • domain assumption The reported benchmark numbers were obtained under a fair protocol matching prior methods.
    No code, logs, or evaluation scripts are provided, so the test and validation numbers are only self-reported.

how reviews work

0 comments
Cite this review

Pith. "Pith review of DySS: Dynamic Queries and State-Space Learning for Efficient 3D Object Detection from Multi-Camera Videos." pith.science (2026). https://pith.science/paper/4PYWXD5N

@misc{pith2026250610242,
  author       = {Pith},
  title        = {Pith review of: DySS: Dynamic Queries and State-Space Learning for Efficient 3D Object Detection from Multi-Camera Videos},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/4PYWXD5N}},
  note         = {Machine review of arXiv:2506.10242}
}
read the original abstract

Camera-based 3D object detection in Bird's Eye View (BEV) is one of the most important perception tasks in autonomous driving. Earlier methods rely on dense BEV features, which are costly to construct. More recent works explore sparse query-based detection. However, they still require a large number of queries and can become expensive to run when more video frames are used. In this paper, we propose DySS, a novel method that employs state-space learning and dynamic queries. More specifically, DySS leverages a state-space model (SSM) to sequentially process the sampled features over time steps. In order to encourage the model to better capture the underlying motion and correspondence information, we introduce auxiliary tasks of future prediction and masked reconstruction to better train the SSM. The state of the SSM then provides an informative yet efficient summarization of the scene. Based on the state-space learned features, we dynamically update the queries via merge, remove, and split operations, which help maintain a useful, lean set of detection queries throughout the network. Our proposed DySS achieves both superior detection performance and efficient inference. Specifically, on the nuScenes test split, DySS achieves 65.31 NDS and 57.4 mAP, outperforming the latest state of the art. On the val split, DySS achieves 56.2 NDS and 46.2 mAP, as well as a real-time inference speed of 33 FPS.

Figures

Figures reproduced from arXiv: 2506.10242 by the authors.

Figure 1
Figure 1. DySS vs. existing SOTA (using ResNet50 backbone) [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Overview of DySS, our proposed efficient 3D object detector for multi-camera videos. With the learnable sparse pillar queries [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Overview of our proposed state-space feature learning module. The state-space model (SSM) sequentially processes the sampled [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Overview of the state-space adaptive mixing module. [PITH_FULL_IMAGE:figures/full_fig_p004_4.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

43 extracted references · 25 canonical work pages

  1. [1]

    Z-forcing: Training stochastic recurrent networks

    Anirudh Goyal ALIAS PARTH GOY AL, Alessandro Sor- doni, Marc-Alexandre Cˆot´e, Nan Rosemary Ke, and Yoshua Bengio. Z-forcing: Training stochastic recurrent networks. Advances in neural information processing systems , 30,

    work page

  2. [2]

    nuscenes: A multi- modal dataset for autonomous driving

    Holger Caesar, Varun Bankiti, Alex H Lang, Sourabh V ora, Venice Erin Liong, Qiang Xu, Anush Krishnan, Yu Pan, Gi- ancarlo Baldan, and Oscar Beijbom. nuscenes: A multi- modal dataset for autonomous driving. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 11621–11631, 2020. 2, 5, 6, 7

    work page 2020

  3. [3]

    End-to- end object detection with transformers

    Nicolas Carion, Francisco Massa, Gabriel Synnaeve, Nicolas Usunier, Alexander Kirillov, and Sergey Zagoruyko. End-to- end object detection with transformers. In European confer- ence on computer vision, pages 213–229. Springer, 2020. 1, 2

    work page 2020

  4. [4]

    Video mamba suite: State space model as a ver- satile alternative for video understanding

    Guo Chen, Yifei Huang, Jilan Xu, Baoqi Pei, Zhe Chen, Zhiqi Li, Jiahao Wang, Kunchang Li, Tong Lu, and Limin Wang. Video mamba suite: State space model as a ver- satile alternative for video understanding. arXiv preprint arXiv:2403.09626, 2024. 3

    arXiv 2024

  5. [5]

    Adamixer: A fast-converging query-based object detector

    Ziteng Gao, Limin Wang, Bing Han, and Sheng Guo. Adamixer: A fast-converging query-based object detector. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 5364–5373, 2022. 5

    work page 2022

  6. [7]

    Mamba: Linear-time sequence modeling with selective state spaces

    Albert Gu and Tri Dao. Mamba: Linear-time sequence modeling with selective state spaces. arXiv preprint arXiv:2312.00752, 2023. 5

    arXiv 2023

  7. [8]

    Efficiently modeling long sequences with structured state spaces

    Albert Gu, Karan Goel, and Christopher R ´e. Efficiently modeling long sequences with structured state spaces. arXiv preprint arXiv:2111.00396, 2021. 2

    arXiv 2021

  8. [9]

    Deep residual learning for image recognition

    Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. Deep residual learning for image recognition. In Proceed- ings of the IEEE conference on computer vision and pattern recognition, pages 770–778, 2016. 5

    2016

Show all 43 references
  1. [10]

    Bevdet4d: Exploit tempo- ral cues in multi-camera 3d object detection

    Junjie Huang and Guan Huang. Bevdet4d: Exploit tempo- ral cues in multi-camera 3d object detection. arXiv preprint arXiv:2203.17054, 2022. 6

    2022 arXiv

  2. [11]

    Bevpoolv2: A cutting-edge implementation of bevdet toward deployment.arXiv preprint arXiv:2211.17111, 2022

    Junjie Huang and Guan Huang. Bevpoolv2: A cutting-edge implementation of bevdet toward deployment.arXiv preprint arXiv:2211.17111, 2022. 1, 2, 6

    2022 arXiv

  3. [12]

    Bevdet: High-performance multi-camera 3d object de- tection in bird-eye-view

    Junjie Huang, Guan Huang, Zheng Zhu, Yun Ye, and Dalong Du. Bevdet: High-performance multi-camera 3d object de- tection in bird-eye-view. arXiv preprint arXiv:2112.11790,

    arXiv

  4. [13]

    Leveraging vision-centric multi-modal expertise for 3d object detection

    Linyan Huang, Zhiqi Li, Chonghao Sima, Wenhai Wang, Jingdong Wang, Yu Qiao, and Hongyang Li. Leveraging vision-centric multi-modal expertise for 3d object detection. Advances in Neural Information Processing Systems , 36,

  5. [14]

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

    Yanqin Jiang, Li Zhang, Zhenwei Miao, Xiatian Zhu, Jin Gao, Weiming Hu, and Yu-Gang Jiang. Polarformer: Multi- camera 3d object detection with polar transformer. In Pro- ceedings of the AAAI conference on Artificial Intelligence , pages 1042–1050, 2023. 6

    2023

  6. [15]

    Centermask: Real- time anchor-free instance segmentation

    Youngwan Lee and Jongyoul Park. Centermask: Real- time anchor-free instance segmentation. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 13906–13915, 2020. 5, 7

    2020

  7. [16]

    Videomamba: State space model for efficient video understanding

    Kunchang Li, Xinhao Li, Yi Wang, Yinan He, Yali Wang, Limin Wang, and Yu Qiao. Videomamba: State space model for efficient video understanding. arXiv preprint arXiv:2403.06977, 2024. 3

    2024 arXiv

  8. [17]

    Unifying voxel-based representation with transformer for 3d object detection

    Yanwei Li, Yilun Chen, Xiaojuan Qi, Zeming Li, Jian Sun, and Jiaya Jia. Unifying voxel-based representation with transformer for 3d object detection. Advances in Neural In- formation Processing Systems, 35:18442–18455, 2022. 6

    2022

  9. [18]

    Bevstereo: Enhancing depth estimation in multi-view 3d object detection with temporal stereo

    Yinhao Li, Han Bao, Zheng Ge, Jinrong Yang, Jianjian Sun, and Zeming Li. Bevstereo: Enhancing depth estimation in multi-view 3d object detection with temporal stereo. In Pro- ceedings of the AAAI Conference on Artificial Intelligence , pages 1486–1494, 2023. 1, 2, 6

    2023

  10. [19]

    Bevdepth: Acquisition of reliable depth for multi-view 3d object detec- tion

    Yinhao Li, Zheng Ge, Guanyi Yu, Jinrong Yang, Zengran Wang, Yukang Shi, Jianjian Sun, and Zeming Li. Bevdepth: Acquisition of reliable depth for multi-view 3d object detec- tion. In Proceedings of the AAAI Conference on Artificial Intelligence, pages 1477–1485, 2023. 6

    2023

  11. [20]

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

    Zhiqi Li, Wenhai Wang, Hongyang Li, Enze Xie, Chong- hao Sima, Tong Lu, Yu Qiao, and Jifeng Dai. Bevformer: Learning bird’s-eye-view representation from multi-camera images via spatiotemporal transformers. In European con- ference on computer vision, pages 1–18. Springer, 2022...

    2022

  12. [21]

    Bevnext: Reviving dense bev frameworks for 3d object de- tection

    Zhenxin Li, Shiyi Lan, Jose M Alvarez, and Zuxuan Wu. Bevnext: Reviving dense bev frameworks for 3d object de- tection. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages 20113– 20123, 2024. 1, 2

    2024

  13. [22]

    Sparse4d: Multi-view 3d object detec- tion with sparse spatial-temporal fusion

    Xuewu Lin, Tianwei Lin, Zixiang Pei, Lichao Huang, and Zhizhong Su. Sparse4d: Multi-view 3d object detec- tion with sparse spatial-temporal fusion. arXiv preprint arXiv:2211.10581, 2022. 1, 2, 6

    2022 arXiv

  14. [23]

    Sparse4d v2: Recurrent temporal fusion with sparse model

    Xuewu Lin, Tianwei Lin, Zixiang Pei, Lichao Huang, and Zhizhong Su. Sparse4d v2: Recurrent temporal fusion with sparse model. arXiv preprint arXiv:2305.14018, 2023. 1, 2, 6

    2023 arXiv

  15. [24]

    Ray denoising: Depth-aware hard negative sampling for multi- view 3d object detection

    Feng Liu, Tengteng Huang, Qianjing Zhang, Haotian Yao, Chi Zhang, Fang Wan, Qixiang Ye, and Yanzhao Zhou. Ray denoising: Depth-aware hard negative sampling for multi- view 3d object detection. arXiv preprint arXiv:2402.03634, 10, 2024. 6

    2024 arXiv

  16. [25]

    Sparsebev: High-performance sparse 3d object de- tection from multi-camera videos

    Haisong Liu, Yao Teng, Tao Lu, Haiguang Wang, and Limin Wang. Sparsebev: High-performance sparse 3d object de- tection from multi-camera videos. In Proceedings of the IEEE/CVF International Conference on Computer Vision , pages 18580–18590, 2023. 1, 2, 3, 4, 5, 6, 7, 8

    2023

  17. [26]

    Petr: Position embedding transformation for multi-view 3d object detection

    Yingfei Liu, Tiancai Wang, Xiangyu Zhang, and Jian Sun. Petr: Position embedding transformation for multi-view 3d object detection. In European Conference on Computer Vi- sion, pages 531–548. Springer, 2022. 2, 6

    2022

  18. [27]

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

    Yingfei Liu, Junjie Yan, Fan Jia, Shuailin Li, Aqi Gao, Tian- cai Wang, and Xiangyu Zhang. Petrv2: A unified framework for 3d perception from multi-camera images. InProceedings of the IEEE/CVF International Conference on Computer Vi- sion, pages 3262–3272, 2023. 1, 2, 6

    2023

  19. [28]

    Vmamba: Visual state space model

    Yue Liu, Yunjie Tian, Yuzhong Zhao, Hongtian Yu, Lingxi Xie, Yaowei Wang, Qixiang Ye, and Yunfan Liu. Vmamba: Visual state space model. arXiv preprint arXiv:2401.10166,

    arXiv

  20. [29]

    Lion: Linear group rnn for 3d object detection in point clouds

    Zhe Liu, Jinghua Hou, Xinyu Wang, Xiaoqing Ye, Jingdong Wang, Hengshuang Zhao, and Xiang Bai. Lion: Linear group rnn for 3d object detection in point clouds. arXiv preprint arXiv:2407.18232, 2024. 3

    2024 arXiv

  21. [30]

    Is pseudo-lidar needed for monocular 3d object detection? In Proceedings of the IEEE/CVF Inter- national Conference on Computer Vision, pages 3142–3152,

    Dennis Park, Rares Ambrus, Vitor Guizilini, Jie Li, and Adrien Gaidon. Is pseudo-lidar needed for monocular 3d object detection? In Proceedings of the IEEE/CVF Inter- national Conference on Computer Vision, pages 3142–3152,

  22. [31]

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

    Jinhyung Park, Chenfeng Xu, Shijia Yang, Kurt Keutzer, Kris M Kitani, Masayoshi Tomizuka, and Wei Zhan. Time will tell: New outlooks and a baseline for temporal multi- view 3d object detection. InThe Eleventh International Con- ference on Learning Representations, 2022. 1, 2, 6

    2022

  23. [32]

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

    Jonah Philion and Sanja Fidler. Lift, splat, shoot: Encoding images from arbitrary camera rigs by implicitly unprojecting to 3d. In European conference on computer vision , pages 194–210. Springer, 2020. 1, 2

    2020

  24. [33]

    Amixer: Adaptive weight mixing for self-attention free vi- sion transformers

    Yongming Rao, Wenliang Zhao, Jie Zhou, and Jiwen Lu. Amixer: Adaptive weight mixing for self-attention free vi- sion transformers. In European Conference on Computer Vi- sion, pages 50–67. Springer, 2022. 5

    2022

  25. [34]

    Simplified state space layers for sequence modeling

    Jimmy TH Smith, Andrew Warrington, and Scott W Linder- man. Simplified state space layers for sequence modeling. arXiv preprint arXiv:2208.04933, 2022. 2

    2022 arXiv

  26. [35]

    Feedback in imitation learning: The three regimes of covariate shift

    Jonathan Spencer, Sanjiban Choudhury, Arun Venkatraman, Brian Ziebart, and J Andrew Bagnell. Feedback in imitation learning: The three regimes of covariate shift. arXiv preprint arXiv:2102.02872, 2021. 5

    2021 arXiv

  27. [36]

    Sparse r-cnn: End-to-end ob- ject detection with learnable proposals

    Peize Sun, Rufeng Zhang, Yi Jiang, Tao Kong, Chen- feng Xu, Wei Zhan, Masayoshi Tomizuka, Lei Li, Zehuan Yuan, Changhu Wang, et al. Sparse r-cnn: End-to-end ob- ject detection with learnable proposals. In Proceedings of the IEEE/CVF conference on computer vision and pattern re...

    2021

  28. [37]

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

    Shihao Wang, Yingfei Liu, Tiancai Wang, Ying Li, and Xi- angyu Zhang. Exploring object-centric temporal modeling for efficient multi-view 3d object detection. In Proceedings of the IEEE/CVF International Conference on Computer Vi- sion, pages 3621–3631, 2023. 1, 2, 3, 6, 7, 8

    2023

  29. [38]

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

    Yue Wang, Vitor Campagnolo Guizilini, Tianyuan Zhang, Yilun Wang, Hang Zhao, and Justin Solomon. Detr3d: 3d object detection from multi-view images via 3d-to-2d queries. In Conference on Robot Learning, pages 180–191. PMLR, 2022. 1, 2, 6

    2022

  30. [39]

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

    Chenyu Yang, Yuntao Chen, Hao Tian, Chenxin Tao, Xizhou Zhu, Zhaoxiang Zhang, Gao Huang, Hongyang Li, Yu Qiao, Lewei Lu, et al. Bevformer v2: Adapting modern image backbones to bird’s-eye-view recognition via perspective su- pervision. In Computer Vision and Pattern Recognitio...

    2023

  31. [40]

    Futuredepth: Learning to predict the future improves video depth estimation

    Rajeev Yasarla, Manish Kumar Singh, Hong Cai, Yunxiao Shi, Jisoo Jeong, Yinhao Zhu, Shizhong Han, Risheek Gar- repalli, and Fatih Porikli. Futuredepth: Learning to predict the future improves video depth estimation. In Proceedings of the European Conference on Computer Vision, 2024. 5

    2024

  32. [41]

    V oxel mamba: Group-free state space models for point cloud based 3d object detection

    Guowen Zhang, Lue Fan, Chenhang He, Zhen Lei, Zhaox- iang Zhang, and Lei Zhang. V oxel mamba: Group-free state space models for point cloud based 3d object detection. arXiv preprint arXiv:2406.10700, 2024. 3

    2024 arXiv

  33. [42]

    Class-balanced grouping and sampling for point cloud 3d object detection

    Benjin Zhu, Zhengkai Jiang, Xiangxin Zhou, Zeming Li, and Gang Yu. Class-balanced grouping and sampling for point cloud 3d object detection. arXiv preprint arXiv:1908.09492,

    1908 arXiv

  34. [43]

    Vision mamba: Efficient visual representation learning with bidirectional state space model

    Lianghui Zhu, Bencheng Liao, Qian Zhang, Xinlong Wang, Wenyu Liu, and Xinggang Wang. Vision mamba: Efficient visual representation learning with bidirectional state space model. arXiv preprint arXiv:2401.09417, 2024. 3

    2024 arXiv

  35. [44]

    Temporal enhanced training of multi-view 3d object detector via historical object prediction

    Zhuofan Zong, Dongzhi Jiang, Guanglu Song, Zeyue Xue, Jingyong Su, Hongsheng Li, and Yu Liu. Temporal enhanced training of multi-view 3d object detector via historical object prediction. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 3781–379...

    2023

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.