Pith. sign in

REVIEW 3 major objections 6 minor 53 references

RCTrans: Radar-Camera Transformer via Radar Densifier and Sequential Decoder for 3D Object Detection

T0 review · 3 major / 6 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read RCTrans is a query-based radar-camera detector that densifies sparse radar tokens and uses a sequential, pruned transformer decoder to reach new state-of-the-art 3D detection results on nuScenes.

desk verdict Solid radar-camera fusion contribution with a genuinely useful densifier and sequential decoder, but the headline SOTA claim rests on quoted baselines and could evaporate under controlled re-runs. read the letter →

arxiv 2412.12799 v1 pith:B6TATWZO submitted 2024-12-17 cs.CV cs.AI

Yiheng Li , Yang Yang , Zhen Lei This is my paper · ORCID
classification cs.CVcs.AI
keywords radar-camerafusion3Dobjectdetectionquery-basedtransformerbird's-eyeviewradarpointclouddensificationsequentialdecodernuScenesmulti-modalperception
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper proposes RCTrans, a query-based detector that fuses camera images and radar point clouds for 3D object detection. It argues that radar's sparsity and noise are best handled before and during fusion: a Radar Dense Encoder fills empty bird's-eye-view grids, and a Pruning Sequential Decoder refines object query positions step by step while using fewer decoder layers at inference. On the nuScenes benchmark the method reports 64.7% NDS and 57.8% mAP on the test set, which the authors call new state-of-the-art radar-camera results, and it also improves 3D tracking and robustness to sensor dropout. The significance, if true, is that radar-camera fusion does not need strict BEV-to-BEV alignment; query-based fusion with densified radar can be more accurate and still fast.

What carries the argument

The load-bearing mechanisms are two. The Radar Dense Encoder (RDE) is a U-Net-style BEV encoder: it downsamples sparse radar BEV features through three scales, applies self-attention with positional embeddings at the smallest scale to fill empty grids adaptively, then upsamples with skip connections to preserve multi-scale detail. The Pruning Sequential Decoder (PSD) updates randomly initialized object queries by first attending to radar tokens with 2D BEV position embeddings, then to image tokens with 3D camera-frustum position embeddings, re-predicting query positions after each layer so the next layer's correspondences are more precise. A pruning strategy trains all six decoder layers but runs only the first three at inference, cutting latency by about 18 ms without dropping NDS.

What would settle it

Re-train the two strongest prior radar-camera detectors on nuScenes val/test with RCTrans's exact training schedule (90 epochs, same augmentations, same inference settings) and compare NDS and mAP. If either baseline reaches or exceeds 64.7% NDS under those conditions, the new-state-of-the-art claim is refuted rather than merely beaten.

Watch

Extended reading notes

Core claim

The central discovery is that both known radar drawbacks have separate fixes that compound. Sparse radar pillars, which leave more than 90% of BEV grids empty, are processed by a downsample-then-upsample encoder with self-attention at the smallest BEV scale and skip connections, so every grid is filled without blurring small objects. Noisy height information from radar is handled by the decoder: object queries first fuse radar tokens through 2D BEV position embeddings, then image tokens through 3D position embeddings, and the query positions are re-projected after every layer so correspondences sharpen progressively. A pruning training strategy, six decoder layers at training and three at inference, both speeds inference and prevents query attention from collapsing onto identical regions. The combination yields the reported 64.7% NDS and 57.8% mAP on nuScenes test with a V2-99 backbone, and ablations attribute +1.3 NDS to the Radar Dense Encoder and +2.2 NDS to the Pruning Sequential Decoder over their immediate baselines.

Load-bearing premise

The state-of-the-art claim assumes the published scores of earlier radar-camera detectors were produced under comparable training and evaluation conditions, because RCTrans quotes those numbers from prior papers rather than re-running them under its own recipe.

Editorial extensions

If this is right

  • Radar-camera fusion can be done in query space instead of aligning BEV grids, so positional offsets between modalities do not need explicit rectification.
  • Densifying radar BEV features before fusion is worth about +1.3 NDS over a SECOND-style encoder, independent of the decoder gain.
  • Sequential per-modality fusion and per-layer query position updates are worth about +2.2 NDS over the immediate baseline.
  • Pruning from six decoder layers at training to three at inference costs nothing in NDS and reduces latency by roughly 18 ms, so the method remains deployment-relevant.
  • The same detector, with velocity-based matching, also improves 3D tracking metrics (AMOTA 59.6) and degrades less than prior methods when camera inputs are dropped.

Reading between the lines

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

  • The authors do not explore adaptive pruning, but the 6-to-3 layer result suggests inference could throttle decoder depth per frame based on query convergence, which would be a natural follow-up.
  • Because RDE targets empty grids rather than radar-specific statistics, it may transfer to other sparse 3D sensors such as 4D radar or low-beam LiDAR without modification.
  • The sequential fusion order is fixed as radar-then-image; a learned or input-dependent ordering might further help when one modality is degraded, though the paper does not test this.
  • The robustness table suggests a practical safety property: with three of six cameras dropped, RCTrans retains most of its car AP, which could matter for sensor-failure deployment; the paper does not analyze failure modes by object class.
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 / 6 minor

Summary. The paper proposes RCTrans, a query-based radar-camera 3D object detector for the nuScenes dataset. It introduces a Radar Dense Encoder that densifies sparse radar BEV features through a downsample-upsample architecture with self-attention and skip connections, and a Pruning Sequential Decoder that alternates radar and image cross-attention layers, updates query positions after each layer, and uses 6 decoder layers at training but only 3 at inference. Experiments report state-of-the-art NDS/mAP on nuScenes test (64.7/57.8 with V2-99), improved performance across multiple image backbones on the val set, 3D tracking results, and robustness to sensor dropout. Ablation studies show stepwise gains for each proposed component.

Significance. If the comparative results are credible, RCTrans is a competitive radar-camera fusion method and among the first query-based approaches to outperform BEV-based radar-camera detectors on nuScenes. The Radar Dense Encoder is a simple and well-motivated remedy for radar sparsity, and the sequential decoder with iterative position refinement is a coherent design. The paper also provides an open-source implementation, which is a valuable contribution. However, the headline state-of-the-art claim rests on a small margin over quoted baselines, and the ablation baseline is a substantially weakened camera stream; these issues must be resolved before the significance can be fully assessed.

major comments (3)
  1. [4.3, Table 2] The claim of a new state of the art rests on a 0.8 NDS margin over RCBEVDet on the nuScenes test set, but all baseline numbers are quoted from prior papers without retraining them under the protocol described in Section 4.2 (90 epochs, batch size 32, cycle LR 4e-4, accumulated radar sweeps, no test-time augmentation). Since no error bars or multi-seed results are reported, the margin could easily be within run-to-run variation. Please provide controlled re-runs of at least RCBEVDet and CRN in your codebase, or report standard deviations over multiple seeds, and explicitly discuss any differences in training settings that may affect comparability.
  2. [4.4, Table 4 and Table 5] The ablations do not cleanly attribute the gains to the proposed modules. The 'camera' baseline is a modified StreamPETR with the spatial alignment module removed and the query position embedding changed, scoring 49.3 NDS versus 54.0 NDS for the original StreamPETR in Table 1; the subsequent fusion and component increments are measured against this weakened variant. Moreover, many increments are 0.1-0.2 NDS (e.g., '+self-attention' 0.1, '+pruning strategy' 0.2), which are likely within noise in the absence of error bars. Please report multiple seeds/standard deviations and add the original StreamPETR as an additional baseline so that the contribution of each component is evaluated against an unmodified reference.
  3. [3.2, Table 4 (sequential structure)] The sequential fusion structure is a core novelty, but the ablation row '+sequential structure' does not specify the baseline architecture or parameter count (e.g., a single concatenated transformer layer with equivalent capacity), and no experiment reverses the fusion order (radar-then-image versus image-then-radar). The claim that this design alleviates elevation ambiguity is therefore not directly tested. Please add a matched-complexity baseline and an order ablation to substantiate the design choice.
minor comments (6)
  1. [Eq. (5) and Section 3.2] The query position embedding PE2d = Φ_ra(Ψ(Rra)) uses Rra, which includes the z coordinate from Eq. (3), but the radar token embedding in Eq. (2) uses only the 2D coordinates (h,w). Clarify that only the x,y components are used for the 2D query embedding, to avoid inconsistency with the stated 2D BEV embedding.
  2. [Section 4.3] The phrase 'giant performance improvement in mA VE' is informal; report the actual margins (e.g., 0.198 vs 0.259) instead.
  3. [Fig. 1] The circle sizes indicate latency, but the exact latency values are not labeled; add numeric labels or a legend with the values so the speed comparison is quantitative.
  4. [Table 4 caption] The symbols △, ▽, and ♢ are used in the 'Modality' rows but are only defined in the text; add definitions to the caption for self-contained reading.
  5. [Section 3.3, Eq. (7)] The loss is written as L(o, ˆo), but the arguments o and ˆo are not defined; define them as the predicted and ground-truth box sets, or write the loss directly in terms of c, b, ĉ, and b̂.
  6. [Section 4.5, Table 6] The meaning of '# of view drops' for radar is ambiguous; specify how many radar sweeps are dropped and how the zero-tensor replacement works for radar.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: RCTrans is an empirical architecture paper evaluated on held-out nuScenes splits, with no fit-then-predict or self-citation chain.

full rationale

The paper's central claim is an empirical performance result on the nuScenes test and validation splits, not a derivation from fitted parameters. The Radar Dense Encoder and Pruning Sequential Decoder are architectural components whose contributions are tested through ablations on the validation set; these ablations compare defined variants and do not redefine the target metric in terms of the components. The pruning schedule (6 training layers, 3 inference layers) is selected from a validation-set sweep in Table 5 and then reported with its measured latency; this is ordinary model selection, not circular reasoning. Baselines in Tables 1-3 are quoted from prior work rather than re-run under identical settings, but that is a comparability/correctness risk, not a circularity: the paper does not derive the baseline numbers from its own model or fit them to its own method. There are no load-bearing self-citations, no imported uniqueness theorems, and no renamed known results. The held-out nuScenes test evaluation provides independent evidence for the stated claim, so the paper is self-contained with respect to circularity concerns despite possible external-validity issues in the SOTA comparison.

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

The ledger shows the central performance claim rests on benchmark correctness, on external baseline comparability, and on a set of unablated hyperparameters. No free constants are fit to produce a 'prediction'; the method is evaluated on held-out nuScenes splits.

free parameters (5)
  • Number of object queries = 900
    Set in Section 4.2; no sensitivity analysis provided, and it directly sets decoder computation and capacity.
  • Radar BEV resolution = 128x128
    Chosen in Section 4.2; this determines how sparse the input grid is and the cost of self-attention in the Radar Dense Encoder.
  • Decoder layers during training and inference = 6 training, 3 inference
    Chosen via the Table 5 sweep; the 6-to-3 operating point is a fitted design choice, and the resulting inference configuration is part of the method.
  • Temporal memory and propagated queries = 512 and 128
    Borrowed from StreamPETR in Section 4.2; not ablated in this paper.
  • Loss weights omega1 and omega2 = not stated
    Defined in Eq. 7 but values are never reported; the classification-regression balance in the total loss is an unspecified free choice.
assumptions (4)
  • domain assumption nuScenes annotations and official evaluation metrics are correct and comparable across methods.
    Every SOTA claim is relative to the nuScenes benchmark (Section 4.1, Tables 1-3).
  • domain assumption Radar elevation is unreliable enough to omit from the radar position embedding.
    Motivates Eq. 2 and the sequential decoder; based on Singh et al. 2023 rather than on a measurement in this paper (Section 3.1).
  • domain assumption Baseline numbers quoted from prior papers were produced under settings comparable to RCTrans.
    The paper does not re-run baselines; comparability is assumed in Section 4.3.
  • domain assumption Temporal fusion using previous frames does not introduce label leakage across the train/val split.
    Standard practice for nuScenes streaming methods; assumed in Section 4.2 when accumulating sweeps and memory.

how reviews work

0 comments
Cite this review

Pith. "Pith review of RCTrans: Radar-Camera Transformer via Radar Densifier and Sequential Decoder for 3D Object Detection." pith.science (2026). https://pith.science/paper/B6TATWZO

@misc{pith2026241212799,
  author       = {Pith},
  title        = {Pith review of: RCTrans: Radar-Camera Transformer via Radar Densifier and Sequential Decoder for 3D Object Detection},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/B6TATWZO}},
  note         = {Machine review of arXiv:2412.12799}
}
read the original abstract

In radar-camera 3D object detection, the radar point clouds are sparse and noisy, which causes difficulties in fusing camera and radar modalities. To solve this, we introduce a novel query-based detection method named Radar-Camera Transformer (RCTrans). Specifically, we first design a Radar Dense Encoder to enrich the sparse valid radar tokens, and then concatenate them with the image tokens. By doing this, we can fully explore the 3D information of each interest region and reduce the interference of empty tokens during the fusing stage. We then design a Pruning Sequential Decoder to predict 3D boxes based on the obtained tokens and random initialized queries. To alleviate the effect of elevation ambiguity in radar point clouds, we gradually locate the position of the object via a sequential fusion structure. It helps to get more precise and flexible correspondences between tokens and queries. A pruning training strategy is adopted in the decoder, which can save much time during inference and inhibit queries from losing their distinctiveness. Extensive experiments on the large-scale nuScenes dataset prove the superiority of our method, and we also achieve new state-of-the-art radar-camera 3D detection results. Our implementation is available at https://github.com/liyih/RCTrans.

Figures

Figures reproduced from arXiv: 2412.12799 by the authors.

Figure 1
Figure 1. Comparison between RCTrans and other camera [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 3
Figure 3. The overall architecture of Radar Dense Encoder. RDE uses a downsample-then-upsample architecture. Self-attention [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figure 4
Figure 4. Qualitative results of RCTrans on nuScenes [PITH_FULL_IMAGE:figures/full_fig_p007_4.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

53 extracted references · 27 canonical work pages

  1. [1]

    Bang, G.; Choi, K.; Kim, J.; Kum, D.; and Choi, J. W. 2024. RadarDistill: Boosting Radar-based Object Detection Performance via Knowledge Distillation from LiDAR Features. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 15491--15500

    2024

  2. [2]

    H.; Vora, S.; Liong, V

    Caesar, H.; Bankiti, V.; Lang, A. H.; Vora, S.; Liong, V. E.; Xu, Q.; Krishnan, A.; Pan, Y.; Baldan, G.; and Beijbom, O. 2020. nuscenes: A multimodal dataset for autonomous driving. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 11621--11631

    2020

  3. [3]

    Carion, N.; Massa, F.; Synnaeve, G.; Usunier, N.; Kirillov, A.; and Zagoruyko, S. 2020. End-to-end object detection with transformers. In European conference on computer vision, 213--229. Springer

    2020

  4. [4]

    R.; and O'Hara, S

    Chaabane, M.; Zhang, P.; Beveridge, J. R.; and O'Hara, S. 2021. Deft: Detection embeddings for tracking. arXiv preprint arXiv:2102.02267

    arXiv 2021

  5. [5]

    Chen, X.; Zhang, T.; Wang, Y.; Wang, Y.; and Zhao, H. 2023. Futr3d: A unified sensor fusion framework for 3d detection. In proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 172--181

    2023

  6. [6]

    Chen, Z.; Li, Z.; Zhang, S.; Fang, L.; Jiang, Q.; and Zhao, F. 2022. Deformable feature aggregation for dynamic multi-modal 3D object detection. In European conference on computer vision, 628--644. Springer

    work page 2022

  7. [7]

    Contributors, M. 2020. MMDetection3D: OpenMMLab next-generation platform for general 3D object detection. https://github.com/open-mmlab/mmdetection3d

    2020

  8. [8]

    Fischer, T.; Yang, Y.-H.; Kumar, S.; Sun, M.; and Yu, F. 2022. Cc-3dt: Panoramic 3d object tracking via cross-camera fusion. arXiv preprint arXiv:2212.01247

    arXiv 2022

Show all 53 references
  1. [9]

    Guizilini, V.; Ambrus, R.; Pillai, S.; Raventos, A.; and Gaidon, A. 2020. 3d packing for self-supervised monocular depth estimation. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2485--2494

    2020

  2. [10]

    He, K.; Zhang, X.; Ren, S.; and Sun, J. 2016. Deep residual learning for image recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition, 770--778

    2016

  3. [11]

    Hu, H.-N.; Yang, Y.-H.; Fischer, T.; Darrell, T.; Yu, F.; and Sun, M. 2022. Monocular quasi-dense 3d object tracking. IEEE Transactions on Pattern Analysis and Machine Intelligence, 45(2): 1992--2008

    2022

  4. [12]

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

    2022 arXiv

  5. [13]

    Huang, J.; and Huang, G. 2022 b . Bevpoolv2: A cutting-edge implementation of bevdet toward deployment. arXiv preprint arXiv:2211.17111

    2022 arXiv

  6. [14]

    Huang, J.; Huang, G.; Zhu, Z.; Ye, Y.; and Du, D. 2021. Bevdet: High-performance multi-camera 3d object detection in bird-eye-view. arXiv preprint arXiv:2112.11790

    2021 arXiv

  7. [15]

    W.; and Kum, D

    Kim, Y.; Kim, S.; Choi, J. W.; and Kum, D. 2023 a . Craft: Camera-radar 3d object detection with spatio-contextual fusion transformer. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 37, 1160--1168

    2023

  8. [16]

    W.; and Kum, D

    Kim, Y.; Shin, J.; Kim, S.; Lee, I.-J.; Choi, J. W.; and Kum, D. 2023 b . Crn: Camera radar net for accurate, robust, efficient 3d perception. In Proceedings of the IEEE/CVF International Conference on Computer Vision, 17615--17626

    2023

  9. [17]

    Lee, Y.; and Park, J. 2020. Centermask: Real-time anchor-free instance segmentation. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 13906--13915

    2020

  10. [18]

    Li, Y.; Bao, H.; Ge, Z.; Yang, J.; Sun, J.; and Li, Z. 2023 a . Bevstereo: Enhancing depth estimation in multi-view 3d object detection with temporal stereo. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 37, 1486--1494

    2023

  11. [19]

    Li, Y.; Ge, Z.; Yu, G.; Yang, J.; Wang, Z.; Shi, Y.; Sun, J.; and Li, Z. 2023 b . Bevdepth: Acquisition of reliable depth for multi-view 3d object detection. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 37, 1477--1485

    2023

  12. [20]

    Li, Z.; Wang, W.; Li, H.; Xie, E.; Sima, C.; Lu, T.; Qiao, Y.; and Dai, J. 2022. Bevformer: Learning bird’s-eye-view representation from multi-camera images via spatiotemporal transformers. In European conference on computer vision, 1--18. Springer

    2022

  13. [21]

    Liang, T.; Xie, H.; Yu, K.; Xia, Z.; Lin, Z.; Wang, Y.; Tang, T.; Wang, B.; and Tang, Z. 2022. Bevfusion: A simple and robust lidar-camera fusion framework. Advances in Neural Information Processing Systems, 35: 10421--10434

    2022

  14. [22]

    Lin, T.-Y.; Goyal, P.; Girshick, R.; He, K.; and Doll \'a r, P. 2017. Focal loss for dense object detection. In Proceedings of the IEEE international conference on computer vision, 2980--2988

    2017

  15. [23]

    Lin, X.; Lin, T.; Pei, Z.; Huang, L.; and Su, Z. 2022. Sparse4d: Multi-view 3d object detection with sparse spatial-temporal fusion. arXiv preprint arXiv:2211.10581

    2022 arXiv

  16. [24]

    Lin, Z.; Liu, Z.; Xia, Z.; Wang, X.; Wang, Y.; Qi, S.; Dong, Y.; Dong, N.; Zhang, L.; and Zhu, C. 2024. RCBEVDet: Radar-camera Fusion in Bird's Eye View for 3D Object Detection. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 14928--14937

    2024

  17. [25]

    Liu, H.; Teng, Y.; Lu, T.; Wang, H.; and Wang, L. 2023 a . Sparsebev: High-performance sparse 3d object detection from multi-camera videos. In Proceedings of the IEEE/CVF International Conference on Computer Vision, 18580--18590

    2023

  18. [26]

    Liu, Y.; Wang, T.; Zhang, X.; and Sun, J. 2022. Petr: Position embedding transformation for multi-view 3d object detection. In European Conference on Computer Vision, 531--548. Springer

    2022

  19. [27]

    Liu, Y.; Yan, J.; Jia, F.; Li, S.; Gao, A.; Wang, T.; and Zhang, X. 2023 b . Petrv2: A unified framework for 3d perception from multi-camera images. In Proceedings of the IEEE/CVF International Conference on Computer Vision, 3262--3272

    2023

  20. [28]

    Liu, Z.; Lin, Y.; Cao, Y.; Hu, H.; Wei, Y.; Zhang, Z.; Lin, S.; and Guo, B. 2021. Swin transformer: Hierarchical vision transformer using shifted windows. In Proceedings of the IEEE/CVF international conference on computer vision, 10012--10022

    2021

  21. [29]

    L.; and Han, S

    Liu, Z.; Tang, H.; Amini, A.; Yang, X.; Mao, H.; Rus, D. L.; and Han, S. 2023 c . Bevfusion: Multi-task multi-sensor fusion with unified bird's-eye view representation. In 2023 IEEE international conference on robotics and automation (ICRA), 2774--2781. IEEE

    2023

  22. [30]

    Long, Y.; Kumar, A.; Morris, D.; Liu, X.; Castro, M.; and Chakravarty, P. 2023. RADIANT: Radar-image association network for 3D object detection. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 37, 1808--1816

    2023

  23. [31]

    Mao, J.; Shi, S.; Wang, X.; and Li, H. 2023. 3D object detection for autonomous driving: A comprehensive survey. International Journal of Computer Vision, 131(8): 1909--1963

    2023

  24. [32]

    Nabati, R.; and Qi, H. 2021. Centerfusion: Center-based radar and camera fusion for 3d object detection. In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, 1527--1536

    2021

  25. [33]

    M.; Tomizuka, M.; and Zhan, W

    Park, J.; Xu, C.; Yang, S.; Keutzer, K.; Kitani, K. M.; Tomizuka, M.; and Zhan, W. 2022. Time will tell: New outlooks and a baseline for temporal multi-view 3d object detection. In The Eleventh International Conference on Learning Representations

    2022

  26. [34]

    Philion, J.; and Fidler, S. 2020. Lift, splat, shoot: Encoding images from arbitrary camera rigs by implicitly unprojecting to 3d. In Computer Vision--ECCV 2020: 16th European Conference, Glasgow, UK, August 23--28, 2020, Proceedings, Part XIV 16, 194--210. Springer

    2020

  27. [35]

    Ronneberger, O.; Fischer, P.; and Brox, T. 2015. U-net: Convolutional networks for biomedical image segmentation. In Medical image computing and computer-assisted intervention--MICCAI 2015: 18th international conference, Munich, Germany, October 5-9, 2015, proceedings, part II...

    2015

  28. [36]

    D.; Ba, Y.; Sarker, A.; Zhang, H.; Kadambi, A.; Soatto, S.; Srivastava, M.; and Wong, A

    Singh, A. D.; Ba, Y.; Sarker, A.; Zhang, H.; Kadambi, A.; Soatto, S.; Srivastava, M.; and Wong, A. 2023. Depth estimation from camera image and mmwave radar point cloud. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 9275--9285

    2023

  29. [37]

    Song, Z.; Yang, L.; Xu, S.; Liu, L.; Xu, D.; Jia, C.; Jia, F.; and Wang, L. 2024. Graphbev: Towards robust bev feature alignment for multi-modal 3d object detection. arXiv preprint arXiv:2403.11848

    2024 arXiv

  30. [38]

    Svenningsson, P.; Fioranelli, F.; and Yarovoy, A. 2021. Radar-pointgnn: Graph based object recognition for unstructured radar point-cloud data. In 2021 IEEE Radar Conference (RadarConf21), 1--6. IEEE

    2021

  31. [39]

    o hler, D.; Niederl \

    Ulrich, M.; Braun, S.; K \"o hler, D.; Niederl \"o hner, D.; Faion, F.; Gl \"a ser, C.; and Blume, H. 2022. Improved Orientation Estimation and Detection with Hybrid Object Detection Networks for Automotive Radar. In arXiv preprint arXiv:2205.02111

    2022 arXiv

  32. [40]

    Wang, S.; Liu, Y.; Wang, T.; Li, Y.; and Zhang, X. 2023. Exploring object-centric temporal modeling for efficient multi-view 3d object detection. In Proceedings of the IEEE/CVF International Conference on Computer Vision, 3621--3631

    2023

  33. [41]

    C.; Zhang, T.; Wang, Y.; Zhao, H.; and Solomon, J

    Wang, Y.; Guizilini, V. C.; Zhang, T.; Wang, Y.; Zhao, H.; and Solomon, J. 2022 a . Detr3d: 3d object detection from multi-view images via 3d-to-2d queries. In Conference on Robot Learning, 180--191. PMLR

    2022

  34. [42]

    Wang, Y.; Zhang, X.; Yang, T.; and Sun, J. 2022 b . Anchor detr: Query design for transformer-based detector. In Proceedings of the AAAI conference on artificial intelligence, volume 36, 2567--2575

    2022

  35. [43]

    Wu, Z.; Chen, G.; Gan, Y.; Wang, L.; and Pu, J. 2023. Mvfusion: Multi-view 3d object detection with semantic-aligned radar and camera fusion. In 2023 IEEE International Conference on Robotics and Automation (ICRA), 2766--2773. IEEE

    2023

  36. [44]

    Xiong, K.; Gong, S.; Ye, X.; Tan, X.; Wan, J.; Ding, E.; Wang, J.; and Bai, X. 2023. Cape: Camera view position embedding for multi-view 3d object detection. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 21570--21579

    2023

  37. [45]

    Yan, J.; Liu, Y.; Sun, J.; Jia, F.; Li, S.; Wang, T.; and Zhang, X. 2023. Cross modal transformer: Towards fast and robust 3d object detection. In Proceedings of the IEEE/CVF International Conference on Computer Vision, 18268--18278

    2023

  38. [46]

    Yan, Y.; Mao, Y.; and Li, B. 2018. Second: Sparsely embedded convolutional detection. Sensors, 18(10): 3337

    2018

  39. [47]

    Yang, B.; Guo, R.; Liang, M.; Casas, S.; and Urtasun, R. 2020. Radarnet: Exploiting radar for robust perception of dynamic objects. In Computer Vision--ECCV 2020: 16th European Conference, Glasgow, UK, August 23--28, 2020, Proceedings, Part XVIII 16, 496--512. Springer

    2020

  40. [48]

    Yin, T.; Zhou, X.; and Krahenbuhl, P. 2021. Center-based 3d object detection and tracking. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 11784--11793

    2021

  41. [49]

    Zheng, L.; Li, S.; Tan, B.; Yang, L.; Chen, S.; Huang, L.; Bai, J.; Zhu, X.; and Ma, Z. 2023. Rcfusion: Fusing 4-d radar and camera with bird’s-eye view features for 3-d object detection. IEEE Transactions on Instrumentation and Measurement, 72: 1--14

    2023

  42. [50]

    Zhou, D.; Kang, B.; Jin, X.; Yang, L.; Lian, X.; Jiang, Z.; Hou, Q.; and Feng, J. 2021. Deepvit: Towards deeper vision transformer. arXiv preprint arXiv:2103.11886

    2021 arXiv

  43. [51]

    Zhou, T.; Chen, J.; Shi, Y.; Jiang, K.; Yang, M.; and Yang, D. 2023. Bridging the view disparity between radar and camera features for multi-modal fusion 3d object detection. IEEE Transactions on Intelligent Vehicles, 8(2): 1523--1535

    2023

  44. [52]

    , " * write output.state after.block = add.period write newline

    ENTRY address archivePrefix author booktitle chapter edition editor eid eprint howpublished institution isbn journal key month note number organization pages publisher school series title type volume year label extra.label sort.label short.list INTEGERS output.state before.all...

  45. [53]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 gl...

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.