Pith. sign in

REVIEW 3 major objections 6 minor 37 references

RbFT-Net: Rectify-Before-Fuse Temporal Radar Anchors for 4D Radar-Camera Depth Completion

T0 review · 3 major / 6 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read Treating accumulated radar as noisy anchors and correcting them with image context before propagation yields denser, more accurate depth than propagating raw returns.

desk verdict Genuinely new rectify-before-fuse design for radar-camera depth completion, well-supported by ablations, but with an overstated iRMSE claim and an unexamined temporal-rectification mechanism. read the letter →

arxiv 2608.13102 v1 pith:DCPASG5K submitted 2026-08-13 cs.CV

Wentao Zhao , Shouxuan Wu , Yongtao Cen , Tianchen Deng , Yuyang Zhang , Jingchuan Wang This is my paper · ORCID
classification cs.CV
keywords 4Dradardepthcompletionradar-camerafusiontemporalaccumulationanchorrectificationreliability-awarepropagationmulti-frameestimationautonomousdriving
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

This paper argues that the reason multi-frame radar often fails to improve depth completion is that accumulated radar returns are treated as accurate measurements, so their errors spread into large image regions during propagation. RbFT-Net instead treats each accumulated radar point as a noisy anchor candidate, corrects its image-plane location and metric depth using local image context, and estimates a pointwise reliability score before any propagation. On the ZJU-4DRadarCam benchmark and a newly collected radar-camera-LiDAR dataset, the resulting predictions beat all evaluated independent radar-camera methods and match or exceed a plug-in configuration that relies on an auxiliary monocular depth model. If the claim holds, dense metric depth can be obtained from radar and RGB alone, with fewer parameters and real-time throughput.

What carries the argument

The load-bearing object is the rectified temporal anchor set, where each raw radar projection is converted into a corrected location, corrected metric depth, reliability score, and anchor feature. The mechanism is the image-conditioned rectification module, which uses a radar query sampler, radar-centered attention, and multi-scale fusion to read local visual context around each candidate, followed by reliability-aware anchor propagation that selects compatible anchors and weights them by reliability before the BP-Net-style MFN-CSPN++ refinement. This machinery is what turns dense but unreliable temporal evidence into sparse but trustworthy metric anchors that can be safely diffused over the image.

What would settle it

Run a sequence with large ego-motion or known calibration misalignment and compare raw projected anchor locations against rectified locations using LiDAR ground truth: if the rectified anchors are not systematically closer to the true returns, or if high-reliability anchors do not have lower depth error than low-reliability anchors, the core premise fails.

Watch

Extended reading notes

Core claim

The central claim is that rectification before fusion, rather than denser accumulation or a stronger backbone, is what unlocks temporal radar cues. The paper projects several radar frames into the current image without ego-motion compensation, forming a set of anchor candidates. An image-conditioned module predicts for each anchor a spatial offset, a depth residual, and a reliability score, producing rectified anchors with corrected positions and depths. A propagation head then scores nearby anchors by target-anchor compatibility, keeps the top four, and aggregates them with reliability-modulated weights before multi-modal fusion and refinement. On ZJU-4DRadarCam, the method achieves the best results among independent methods across all reported metrics, including an iMAE of 6.8 versus 10.4 for the next-best independent method, and remains competitive with the plug-in TacoDepth configuration without using an auxiliary monocular depth model.

Load-bearing premise

The rectification module must recover true image positions and depths from local image context alone, because radar points are projected into the current frame without ego-motion compensation; if misprojection or temporal misalignment is too large for local image features to disambiguate, the rectify-before-fuse premise collapses.

Editorial extensions

If this is right

  • A dense depth map can be produced from radar plus RGB alone at 44.88 FPS with 44.2 million parameters, about one-third the parameter count of the compared plug-in pipelines, while matching or beating them on the tested benchmarks.
  • Temporal accumulation helps only when errors are corrected beforehand: RbFT-Net's RMSE drops from 2987.3 mm to 2740.8 mm as the number of frames rises from one to five, whereas direct-accumulation baselines gain little from additional frames.
  • Predicted anchor reliability is a usable gate: binned statistics show that anchor depth error decreases as predicted reliability increases, and that high-reliability anchors are selected more often during learned propagation.
  • The method transfers across sensing platforms: zero-shot from ZJU-4DRadarCam to a new radar-camera-LiDAR rig reduces MAE by 22.6% relative to the best compared baseline, and fine-tuning on 10% of the target-domain training set improves all metrics.
  • Using seven frames yields slightly worse results than five, indicating that the optimal accumulation window is a trade-off between observation density and temporal noise.

Reading between the lines

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

  • The authors leave implicit that the same rectify-before-fuse discipline could apply to other sparse temporal sensors, such as automotive radar without elevation, ultrasonic arrays, or noisy depth from structure-from-motion, where accumulation adds misalignment faster than it adds information.
  • A natural extension would be an adaptive accumulation window: the reported degradation at seven frames suggests the optimal window depends on scene dynamics, so a learned per-frame weight or a stop-fusing decision could push performance further.
  • Because rectification reads only local image context, failure cases are plausible under large inter-frame motion or in textureless regions; explicit ego-motion compensation before rectification is a complementary fix that the paper does not test.
  • The reliability score could plausibly be turned into a calibrated probability, for example by adding a temperature-scaled confidence loss, which would make the confidence head usable for downstream uncertainty-aware planning.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 6 minor

Summary. The paper proposes RbFT-Net, a 4D radar-camera dense depth completion framework built around a rectify-before-fuse principle. Multi-frame radar returns are accumulated by direct projection into the current view (Eq. (1)) and treated as noisy anchor candidates. An image-conditioned module estimates per-anchor spatial offsets, depth residuals, and reliabilities (Eqs. (14)-(17)); a reliability-aware propagation module then selects the K_s most compatible anchors for each target and aggregates them with reliability weights (Eqs. (18)-(23)) before MFN-CSPN++ refinement. The method is trained end-to-end with LiDAR depth supervision. Experiments on ZJU-4DRadarCam and a newly collected dataset, including zero-shot and limited-data adaptation, show that RbFT-Net outperforms the evaluated independent methods and is competitive with plug-in methods that use an auxiliary monocular depth model.

Significance. If the results are robust, RbFT-Net provides a compact alternative to plug-in pipelines that require an auxiliary monocular model, and its explicit modeling of anchor reliability is a sensible response to the noise and sparsity of 4D radar. The paper's strengths include a thorough component ablation (Table 4), a temporal-frame analysis (Table 5), and cross-platform evaluation on a new dataset, which are uncommon in this line of work. The principal caveats are that the state-of-the-art claim rests on an unequal input configuration (five radar frames for RbFT-Net versus one for competitors) and that the geometric rectification is not cleanly separated from reliability-based selection.

major comments (3)
  1. [Comparison with State-of-the-Art Methods (Table 1)] The text states that RbFT-Net reduces iRMSE by 'over 25%' relative to the strongest independent baseline. According to Table 1 (0-70 m), the best independent baseline iRMSE is 21.0 (BP-Net) and RbFT-Net achieves 16.8, which is a 20% reduction; the 'over 25%' figure is only reached if the comparison is made against SomeDepth (22.8), which is not the metric-specific best. Please correct the reported improvement or explicitly state the baseline used for each metric.
  2. [Comparison with State-of-the-Art Methods (Tables 1 and 5)] The headline comparison in Table 1 gives RbFT-Net five accumulated radar frames while all competitors use a single frame. Table 5 shows that RbFT-Net with one frame reaches RMSE 2987.3 at 0-70 m, which is worse than TacoDepth-independent (2906.3) and SomeDepth (2946.9) reported in Table 1. The claim that 'RbFT-Net consistently achieves the best performance among independent methods across all metrics and evaluation ranges' is therefore only true for the five-frame configuration. Please report full-metric results for the one-frame setting and qualify the state-of-the-art claim accordingly.
  3. [Image-Conditioned Radar Anchor Rectification (Eq. (1))] Because Eq. (1) projects all T frames with the current extrinsics without ego-motion compensation, and the anchors are processed as an unordered set without temporal-index encoding, the rectification module must implicitly recover motion-induced misprojections. The paper does not provide a diagnostic relating the predicted spatial offsets (Delta p_i) to actual projection errors caused by ego-motion, nor does it include an ablation with ego-compensated input. Without such evidence, it is unclear whether the multi-frame gains come from geometrically correcting misprojected anchors or primarily from the confidence head and top-K selection, which suppress unreliable returns rather than rectify them. This distinction is central to the 'rectify-before-fuse' contribution and should be addressed with targeted experiments.
minor comments (6)
  1. [Training Objectives] Eq. (25) introduces L_anchor, L_conf, and L_prop, but their definitions are only promised in the supplementary material. At least a one-sentence definition of each loss should be included in the main text, since these objectives are central to the method.
  2. [Experimental Setup / Compared Methods] The method listed as 'SomeDepth' in Table 1 is attributed to Hou and Ohtsuki (2026), but the name and architecture are not explained; please use a more descriptive reference or clarify the method.
  3. [Method / Overview] The notation F^t_rad in Eq. (24) is not defined in the main text; please define it as the feature derived from the current-frame radar returns.
  4. [Experimental Setup / Evaluation Metrics] iMAE and iRMSE are reported in units of 1/km; please state this unit explicitly in the table captions or headers to avoid misreading of the absolute values.
  5. [Table 4] The row labels in Table 4 are difficult to parse in the provided text version; please ensure that the checkmarks clearly indicate which of the Spatial, Depth, and Confidence components are enabled for each row.
  6. [Datasets] The newly collected dataset is introduced without a description of its size, sensor platform, or scene distribution in the main text; brief statistics would help readers assess the cross-platform evaluation.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the rectification and propagation claims are empirically evaluated against external benchmarks and do not reduce by construction to their inputs.

full rationale

The paper's central claim, that rectifying accumulated multi-frame radar anchors with image-conditioned reliability estimation before propagation improves dense depth, is an empirical claim supported by experiments on the external ZJU-4DRadarCam benchmark and a newly collected dataset. No equation in the method section reduces by construction to a fitted quantity or to the final prediction: the rectified anchor location and depth are residual updates (Eq. 16) whose offsets and residuals are predicted from image and radar features, and the final depth map is trained with LiDAR-projected supervision through L_depth and the auxiliary losses in Eq. 25. The ablations in Table 4 isolate spatial rectification, depth rectification, confidence, and learned propagation, and the gains are reported as measured differences rather than as consequences of the definitions. The choice of five accumulated radar frames is a hyperparameter selected from Table 5, which is tuning rather than circularity. The paper cites RaCalNet by overlapping authors, but only as related work noting prior recalibration efforts, and this citation is not load-bearing for the proposed framework. The newly collected dataset is self-collected, but the main quantitative comparison also uses the public ZJU-4DRadarCam dataset, and the paper additionally reports zero-shot and limited-data adaptation results. These are experimental and materials considerations, not circular derivation steps. Overall, the claims are self-contained against external benchmarks and no circular step can be exhibited with a specific equation or fitted-parameter-as-prediction reduction.

Assumptions & free parameters 7 free parameters · 5 assumptions · 0 invented entities

No new physical entities, forces, particles, or conserved quantities are introduced. The rectified anchor, reliability score, and compatibility score are learned representations within the network, not external entities with falsifiable handles. The free parameters are hyperparameters (T, K, Ks, loss weights, resolution, M, S) that affect the empirical results but are not derived from theory.

free parameters (7)
  • Number of accumulated radar frames T = 5 (chosen from {1,2,3,5,7})
    Selected as default after observing in Table 5 that RMSE on ZJU-4DRadarCam improves from 2987.3 mm (T=1) to 2740.8 mm (T=5) and slightly worsens at T=7 (2760.2 mm). This is a benchmark-driven hyperparameter choice, not a derived quantity.
  • Number of selected anchors per target Ks = 4
    Set to 4 in all experiments (reliability-aware propagation section). No ablation or sensitivity analysis for Ks is provided.
  • Number of retrieved neighbor anchors K = 8
    K=8 rectified anchors are retrieved per target location in all experiments (Implementation Details).
  • Loss weights lambda_d, lambda_a, lambda_c, lambda_p = not reported in main text
    The total loss L is a weighted sum of four terms (Eq. 25). The actual weights are deferred to the supplementary material, so the balance is a free fitting choice not auditable from the preprint.
  • Input resolution = 288x864 (ZJU), 288x832 (new dataset)
    Different per dataset and not ablated. Resolution affects both accuracy and the FPS comparison.
  • Number of radar-conditioned sampling offsets M = not reported
    RQS samples M offsets per scale (Eq. 6). M is an architecture hyperparameter not ablated or reported in the main text.
  • Number of feature scales S = not reported
    MAF fuses S scales (Eq. 14). The value of S is not stated in the main text.
assumptions (5)
  • domain assumption LiDAR-projected depth maps are accurate ground truth for supervision and evaluation.
    Training uses L_depth on LiDAR-projected depth (Section 'Training Objectives'), and evaluation metrics use valid ground-truth pixels. Any systematic LiDAR-camera misalignment would bias results.
  • domain assumption Radar-camera extrinsic calibration and camera intrinsics are accurate enough that initial projections in Eq. (1) are plausible anchors.
    Eq. (1) directly projects radar returns using calibrated extrinsics and intrinsics. If calibration were grossly wrong, the image-conditioned rectification would have no way to recover.
  • ad hoc to paper Local image context around each initial projection contains enough visual evidence to correct misprojections and depth errors.
    The rectification module samples features near p_i plus learned offsets (RQS, Eqs. 6-7). This is the core modeling premise of rectify-before-fuse and is not proven; large projection errors or textureless regions would break it.
  • ad hoc to paper Temporal misalignment from omitting ego-motion compensation is learnable from image evidence.
    Eq. (1) accumulates radar frames without ego-motion compensation; the network must implicitly handle ego and object motion. The paper does not quantify how much misalignment the rectification can absorb.
  • standard math Standard deep learning building blocks (cross-attention, KNN, CSPN++) behave as described and are appropriate inductive biases.
    The architecture relies on these established operations; no new mathematical guarantees are given.

how reviews work

0 comments
Cite this review

Pith. "Pith review of RbFT-Net: Rectify-Before-Fuse Temporal Radar Anchors for 4D Radar-Camera Depth Completion." pith.science (2026). https://pith.science/paper/DCPASG5K

@misc{pith2026260813102,
  author       = {Pith},
  title        = {Pith review of: RbFT-Net: Rectify-Before-Fuse Temporal Radar Anchors for 4D Radar-Camera Depth Completion},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/DCPASG5K}},
  note         = {Machine review of arXiv:2608.13102}
}
read the original abstract

Dense metric depth prediction from cameras and millimeter-wave radar offers a cost-effective sensing solution for autonomous systems. However, radar measurements are inherently sparse and susceptible to clutter, multipath reflections, and projection errors. While aggregating multiple radar frames provides denser metric cues, it also introduces temporal misalignment and dynamic-object interference. Directly propagating such unreliable measurements can therefore corrupt large regions of the predicted depth map. To address this issue, we propose RbFT-Net, an end-to-end rectify-before-fuse framework for multi-frame 4D radar-camera depth completion. Rather than assuming accumulated radar returns to be accurate, RbFT-Net treats them as noisy temporal anchor candidates. An image-conditioned rectification module jointly corrects their image-plane locations and metric depths while estimating pointwise reliability. The rectified anchors are then selectively propagated before high-level multi-modal fusion, suppressing the influence of unreliable measurements. Experiments on ZJU-4DRadarCam and a newly collected 4D radar-camera-LiDAR dataset show that RbFT-Net consistently outperforms the evaluated independent radar-camera methods and remains competitive with plug-in pipelines using auxiliary monocular depth models. Cross-platform evaluation and component analyses further support the effectiveness of the proposed rectification and reliability-aware propagation strategy.

Figures

Figures reproduced from arXiv: 2608.13102 by the authors.

Figure 1
Figure 1. Overview of RbFT-Net. Directly accumulated multi-frame 4D radar returns are treated as noisy temporal anchor [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Image-conditioned radar anchor rectification. RQS and RCA aggregate local visual evidence around each radar [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Reliability-aware anchor propagation. For each tar [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Qualitative comparison on ZJU-4DRadarCam. [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]
Figure 5
Figure 5. Figure 5: Reliability analysis of rectified radar anchors. [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]
Figure 6
Figure 6. Figure 6: Visualization of learned target–anchor associations. [PITH_FULL_IMAGE:figures/full_fig_p007_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

37 extracted references · 30 canonical work pages

  1. [1]

    2024 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) , pages=

    Cafnet: A confidence-driven framework for radar camera depth estimation , author=. 2024 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) , pages=. 2024 , organization=

    work page 2024

  2. [2]

    2025 IEEE/CVF Winter Conference on Applications of Computer Vision (WACV) , pages=

    Get-up: Geometric-aware depth estimation with radar points upsampling , author=. 2025 IEEE/CVF Winter Conference on Applications of Computer Vision (WACV) , pages=. 2025 , organization=

    work page 2025

  3. [3]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    Depth estimation from camera image and mmwave radar point cloud , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    work page

  4. [4]

    2024 IEEE International Conference on Robotics and Automation (ICRA) , pages=

    Radarcam-depth: Radar-camera fusion for depth estimation with learned metric scale , author=. 2024 IEEE International Conference on Robotics and Automation (ICRA) , pages=. 2024 , organization=

    work page 2024

  5. [5]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    Tacodepth: Towards efficient radar-camera depth estimation with one-stage fusion , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    work page

  6. [6]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    Radar-camera pixel depth association for depth completion , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    work page

  7. [7]

    ICASSP 2025-2025 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) , pages=

    Lircdepth: Lightweight radar-camera depth estimation via knowledge distillation and uncertainty guidance , author=. ICASSP 2025-2025 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) , pages=. 2025 , organization=

    work page 2025

  8. [8]

    XD-RCDepth: Lightweight Radar-Camera Depth Estimation with Explainability-Aligned and Distribution-Aware Distillation

    XD-RCDepth: Lightweight Radar-Camera Depth Estimation with Explainability-Aligned and Distribution-Aware Distillation , author=. arXiv preprint arXiv:2510.13565 , year=

    work page Pith review arXiv

Show all 37 references
  1. [9]

    IEEE Robotics and Automation Letters , year=

    JustDepth: Real-Time Radar-Camera Depth Estimation With Single-Scan LiDAR Supervision , author=. IEEE Robotics and Automation Letters , year=

  2. [10]

    European Conference on Computer Vision , pages=

    Sparse beats dense: Rethinking supervision in radar-camera depth completion , author=. European Conference on Computer Vision , pages=. 2024 , organization=

    2024

  3. [11]

    arXiv preprint arXiv:2506.15560 , year=

    RaCalNet: Radar Calibration Network for Sparse-Supervised Metric Depth Estimation , author=. arXiv preprint arXiv:2506.15560 , year=

    arXiv

  4. [12]

    IEEE Transactions on Circuits and Systems for Video Technology , volume=

    CSPN: A category-specific processing network for low-light image enhancement , author=. IEEE Transactions on Circuits and Systems for Video Technology , volume=. 2024 , publisher=

    2024

  5. [13]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    Bidirectional projection network for cross dimension scene understanding , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  6. [14]

    European conference on computer vision , pages=

    Non-local spatial propagation network for depth completion , author=. European conference on computer vision , pages=. 2020 , organization=

    2020

  7. [15]

    2021 IEEE International Conference on Image Processing (ICIP) , pages=

    Depth estimation from monocular images and sparse radar using deep ordinal regression network , author=. 2021 IEEE International Conference on Image Processing (ICIP) , pages=. 2021 , organization=

    2021

  8. [16]

    Displays , volume=

    Self-supervised deep monocular visual odometry and depth estimation with observation variation , author=. Displays , volume=. 2023 , publisher=

    2023

  9. [17]

    Proceedings of the IEEE/CVF international conference on computer vision , pages=

    Digging into self-supervised monocular depth estimation , author=. Proceedings of the IEEE/CVF international conference on computer vision , pages=

  10. [18]

    1--a model zoo for robust monocular relative depth estimation , author=

    Midas v3. 1--a model zoo for robust monocular relative depth estimation , author=. arXiv preprint arXiv:2307.14460 , year=

    arXiv

  11. [19]

    Advances in Neural Information Processing Systems , volume=

    Depth anything v2 , author=. Advances in Neural Information Processing Systems , volume=

  12. [20]

    arXiv preprint arXiv:2302.12288 , year=

    ZoeDepth: Zero-shot Transfer by Combining Relative and Metric Depth , author=. arXiv preprint arXiv:2302.12288 , year=

    arXiv

  13. [21]

    Proceedings of the IEEE/CVF International Conference on Computer Vision , year=

    Metric3D: Towards Zero-shot Metric 3D Prediction from A Single Image , author=. Proceedings of the IEEE/CVF International Conference on Computer Vision , year=

  14. [22]

    2020 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) , pages=

    Depth estimation from monocular images and sparse radar data , author=. 2020 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) , pages=. 2020 , organization=

    2020

  15. [23]

    Neurocomputing , volume=

    RCDformer: Transformer-based dense depth estimation by sparse radar and camera , author=. Neurocomputing , volume=. 2024 , publisher=

    2024

  16. [24]

    2025 IEEE International Conference on Robotics and Automation (ICRA) , pages=

    Structure-Aware Radar-Camera Depth Estimation , author=. 2025 IEEE International Conference on Robotics and Automation (ICRA) , pages=. 2025 , organization=

    2025

  17. [25]

    arXiv preprint arXiv:2503.17182 , year=

    Radar-guided polynomial fitting for metric depth estimation , author=. arXiv preprint arXiv:2503.17182 , year=

    arXiv

  18. [26]

    CAAI Transactions on Intelligence Technology , year=

    Uno: Unified self-supervised monocular odometry for platform-agnostic deployment , author=. CAAI Transactions on Intelligence Technology , year=

  19. [27]

    Advances in Neural Information Processing Systems , volume=

    K-radar: 4d radar object detection for autonomous driving in various weather conditions , author=. Advances in Neural Information Processing Systems , volume=

  20. [28]

    arXiv preprint arXiv:2605.11840 , year=

    Selection, Not Fusion: Radar-Modulated State Space Models for Radar-Camera Depth Estimation , author=. arXiv preprint arXiv:2605.11840 , year=

    arXiv

  21. [29]

    Sensors , volume=

    Expanding sparse radar depth based on joint bilateral filter for radar-guided monocular depth estimation , author=. Sensors , volume=. 2024 , publisher=

    2024

  22. [30]

    2021 International Conference on 3D Vision (3DV) , pages=

    R4dyn: Exploring radar for self-supervised monocular depth estimation of dynamic scenes , author=. 2021 International Conference on 3D Vision (3DV) , pages=. 2021 , organization=

    2021

  23. [31]

    Proceedings of the AAAI conference on artificial intelligence , volume=

    Cspn++: Learning context and resource aware convolutional spatial propagation networks for depth completion , author=. Proceedings of the AAAI conference on artificial intelligence , volume=

  24. [32]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    Bilateral propagation network for depth completion , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  25. [33]

    Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

    The temporal opportunist: Self-supervised multi-frame monocular depth , author=. Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

  26. [34]

    IEEE Robotics and Automation Letters , year=

    Manydepth2: Motion-aware self-supervised monocular depth estimation in dynamic scenes , author=. IEEE Robotics and Automation Letters , year=

  27. [35]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    Multi-Frame Self-Supervised Depth with Transformers , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  28. [36]

    Proceedings of the IEEE/CVF international conference on computer vision , pages=

    Vision transformers for dense prediction , author=. Proceedings of the IEEE/CVF international conference on computer vision , pages=

  29. [37]

    2025 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) , pages=

    A 4D Radar Camera Extrinsic Calibration Tool Based on 3D Uncertainty Perspective N Points , author=. 2025 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) , pages=. 2025 , organization=

    2025

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.