Pith. sign in

REVIEW 4 major objections 6 minor 5 cited by

Metric Anything claims that metric depth estimation finally scales: roughly 20M heterogeneous image-depth pairs produce steady accuracy gains and transfer across depth completion, monocular depth, camera calibration, 3D reconstruction, and

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

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

T0 review · deepseek-v4-flash

2026-08-03 06:43 UTC pith:NUPJRV7L

load-bearing objection Promising scaling idea for metric depth, but the metricity of the 20M labels is unverified and one hyperparameter is tuned on an evaluation set; deserves a real referee but needs major fixes. the 4 major comments →

arxiv 2601.22054 v2 pith:NUPJRV7L submitted 2026-01-29 cs.CV cs.AI

MetricAnything: Scaling Metric Depth Pretraining with Noisy Heterogeneous Sources

Baorui Ma , Jiahui Yang , Donglin Di , Xuancheng Zhang , Jianxun Cui , Hao Li , Yan Xie , Wei Chen This is my paper
classification cs.CV cs.AI
keywords metric depth estimationscaling lawssparse promptsheterogeneous 3D datapretrainingdistillationzero-shot generalizationspatial reasoning
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

This paper sets out to prove that metric depth estimation—predicting real physical distances from images—can be scaled like other vision foundation tasks, even when the training data is noisy and heterogeneous. The authors assemble about 20 million image-depth pairs from reconstructed, captured, and rendered 3D sources spanning over 10,000 camera models, and pretrain a single model with randomly masked depth maps as sparse prompts. Their central claim is that this is the first clear scaling trend in the metric depth track: more data consistently improves zero-shot depth accuracy. The same pretrained model, after distillation into a prompt-free student, reports strong results on monocular depth, camera-intrinsics recovery, multi-view 3D reconstruction, and vision-language-action planning, and its encoder improves spatial reasoning in vision-language models. If true, the practical implication is that large-scale 'data soup' pretraining, rather than small hand-curated datasets, is a viable route to general metric perception.

Core claim

The discovery, on the paper's own terms, is that the heterogeneity and noise that kept metric depth data from scaling can be absorbed by a single universal interface: the Sparse Metric Prompt, a randomly masked depth map. With roughly 20M such image-prompt pairs, the pretrained model exhibits a monotonic improvement in zero-shot accuracy as data grows, and the trend carries through distillation to a prompt-free student. The authors take this as evidence that metric depth estimation obeys the same scaling behavior as relative depth and other vision foundation tasks, and that a single pretrained metric representation can serve depth completion, monocular depth, camera calibration, 3D reconstru

What carries the argument

Sparse Metric Prompt—a depth map randomly subsampled to 2,000-40,000 valid pixels per image (about 1% of pixels), then regularized through pixel-wise depth scale alignment and global metric depth recovery against a prior depth prediction so that all data sources share one input format. The prompt is injected through a lightweight conditioned dense-prediction head (about 5% extra parameters), leaving the vision-transformer backbone untouched. Pretraining uses a robust MAE loss that discards the top 20% largest-error pixels per image to tolerate sensor and reconstruction noise. The teacher is distilled into a prompt-free student with a distance-balanced inverse-depth loss, which keeps near-fie

Load-bearing premise

The claim assumes every aggregated source can be converted to per-pixel metric depth from known calibration and that residual errors are pixel-level noise, because the training loss actively discards the largest-error pixels; if a major source carries a systematic scale or calibration bias, the learned metric scale is corrupted instead of cleaned.

What would settle it

Take one large reconstructed source, multiply all its depth values by a fixed 15% scale, retrain at full data scale, and measure zero-shot relative error on several held-out datasets. If the model's error is sharply worse, the robust loss is absorbing source-level bias rather than pixel noise, and the scaling claim would not hold for arbitrary heterogeneous sources. A lighter version: report per-source residual histograms against a high-precision reference and look for a depth-ratio peak tied to a single source.

Watch this falsifier — get emailed when new claim-graph text bears on it.

If this is right

  • Data quality trade-offs shift: a large, noisy, diverse collection can outperform a small clean one for metric depth, so future work can mine existing 3D datasets instead of curating small ones.
  • The pretrained prompt-conditional model is usable out of the box for depth completion and super-resolution across different prompt densities and unseen sensors, including radar.
  • The distilled student inherits metric scale without prompts, enabling monocular metric depth, focal-length recovery, and monocular 3D point-map prediction in camera-unknown settings.
  • The same visual encoder can be reused as a frozen backbone for vision-language models and robot policies, implying metric 3D pretraining is a general spatial representation rather than a task-specific module.
  • Zero-shot behavior on unseen camera configurations, night and rainy conditions, and panoramic/fisheye imagery follows directly from learning heterogeneous sources instead of modeling any one sensor.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The paper does not audit per-source quality; a direct consequence of its argument is that source diversity, not just raw count, drives the scaling curve, and a per-source diagnostic would reveal whether any single reconstruction corpus dominates the gain.
  • Because the student is trained on teacher pseudo-labels from a unified model, the framework implies a self-improving loop—using student outputs as new pseudo-labels and retraining—which the authors do not test but is a natural next step.
  • The success with radar, an unseen and extremely sparse sensor, suggests the same interface may extend to other sparse modalities such as event cameras or time-of-flight arrays, a testable prediction.
  • The vision-language results imply the encoder stores recoverable metric scale; an untested but plausible extension is using it for egocentric distance estimation in embodied navigation, where fine-grained metric errors may matter more than in the benchmarks shown.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 6 minor

Summary. The paper proposes Metric Anything, a pretraining framework for metric depth estimation that aggregates roughly 20M image–depth pairs from reconstructed, captured, and rendered 3D data. The method uses randomly masked depth maps as "Sparse Metric Prompts" to train a depth-conditioned teacher, which is then distilled into a prompt-free student with a distance-balanced inverse-depth loss. The authors claim, for the first time, a clear scaling trend in metric depth pretraining, and report state-of-the-art zero-shot results across depth completion/super-resolution, radar–camera fusion, monocular depth estimation, camera intrinsics recovery, multi-view metric 3D reconstruction, VLA planning, and MLLM spatial reasoning. The paper emphasizes a minimalist, data-driven paradigm with no task-specific prompt engineering.

Significance. If the central claim holds, the paper would establish a valuable scaling route for metric depth estimation, analogous to progress in relative depth. The work is substantial in empirical scope: it aggregates a very large heterogeneous dataset, demonstrates a monotonic trend with data size, and evaluates across many downstream tasks. The paper also ships an open-source release, which is a concrete strength. The main scientific value lies in the scaling claim and in the demonstration that a simple random-mask prompt can unify heterogeneous sources. However, the current evidence for the scaling claim rests on a single benchmark, and at least one hyperparameter appears to have been tuned on an evaluation dataset, which tempers the strength of the reported SOTA results.

major comments (4)
  1. [Sec. 5.7 / Tab. 15 (right) and Tabs. 5–6] The student-loss hyperparameter C is selected on DIODE (Tab. 15, right), and DIODE also appears as an evaluation benchmark in the monocular geometry estimation tables (Tabs. 5 and 6). This contradicts the strict zero-shot protocol claimed in Sec. 4.2.1, where the six datasets are stated to be unused for hyper-parameter tuning. The DIODE rows in Tabs. 5–6 are therefore not trustworthy as zero-shot results, and the average ranks that include them are inflated. Please retune C on a held-out set disjoint from all evaluation benchmarks, or remove DIODE from the evaluation tables.
  2. [Sec. 5.1 / Fig. 2a] The central scaling claim is supported only by δ1 accuracy on Middlebury (Fig. 2a), with no error bars, no multiple training runs, and no description of whether the data fractions are nested or independently sampled. A single benchmark without variance does not establish a 'clear scaling trend' robustly, especially given the noise inherent in heterogeneous data. Please report scaling curves on at least two or three benchmarks, include multiple seeds or confidence intervals, and clarify the subset construction.
  3. [Sec. 3.1 and Eq. (8)] The metric premise of the entire method is asserted but not audited. Section 3.1 includes 'reconstructed 3D data' from SfM, SLAM, MVS, and stereo, whose outputs are not automatically metric; monocular SfM is scale-ambiguous, SLAM can drift, and stereo requires known baselines. The robust MAE loss in Eq. (8) drops only the top-20% per-image largest errors, which cannot correct a coherent per-source scale or offset error. The prompt preparation step (Sec. 3.2) aligns prompts to sampled depth values but does not correct systematic source-level bias in the target G. Please provide a dataset manifest with per-source scale verification, or independent checks (e.g., comparison against LiDAR or known camera trajectories), to substantiate the claim that all 20M pairs are metric along the camera z-axis.
  4. [Sec. 3.3] The prompt-free student is trained entirely on teacher-generated pseudo-labels. This is circular in the sense that any systematic metric error in the teacher (e.g., from the unverified sources above) is inherited by the student. The student results therefore cannot be used as evidence that the aggregated data are metric. This is not a flaw per se, but it means the student's SOTA numbers do not independently validate the metric premise; they only show consistency with the teacher.
minor comments (6)
  1. [Sec. 7.3] Typo: 'MoGo-2' should be 'MoGe-2' (the method referenced in [110]).
  2. [Sec. 5.7] The text refers to 'Eq.14' when describing the balance weight C; the correct reference is Eq. (5).
  3. [Sec. 9] Typo: 'metric depth trick' should be 'metric depth track.'
  4. [Sec. 4.2.3] The term 'boundaris' should be 'boundaries' in the section heading and text.
  5. [Sec. 4.2.6 and Fig. 12] The benchmark is called VSI-Bench in the text and table, but 'VIS Benchmark' appears in Fig. 12; please make the naming consistent.
  6. [References] Reference [36] and [35] appear to be the same paper; please deduplicate.

Circularity Check

0 steps flagged

No significant circularity: the central scaling claim is empirically evaluated on external benchmarks and not forced by construction.

full rationale

Whether the ~20M-pair pretraining yields metric depth is an empirical question, and the paper's own evidence is not constructed from its inputs. The teacher is trained directly on aggregated ground-truth depth maps (Eq. 1, Sec. 3.1) with MAE/SSI-MAGE losses; the sparse prompt is a random mask of the target, so no prediction is fitted to an external quantity and then renamed. The reported scaling trend (Fig. 2a) is evaluated zero-shot on Middlebury for models trained on increasing data fractions, and downstream prompt tasks use held-out datasets (NYUv2, ETH3D, KITTI, nuScenes), so the central numbers are externally anchored. The prompt-free student is distilled from teacher pseudo-labels (Sec. 3.3); this is a self-referential training loop, but the paper's central scaling claim is not derived from the student—it comes from the teacher's zero-shot curve. The only self-citation found, See3D [68], is used for auxiliary PDSA/GMDR prompt alignment; it is not a uniqueness theorem or an unvalidated load-bearing reduction, and the same prompt preparation is at most an implementation detail rather than a step that forces the central result. The Limitations section admits the architecture's own scalability is unexplored, which undercuts the breadth of the scaling claim but is not a circularity.

Axiom & Free-Parameter Ledger

6 free parameters · 5 axioms · 1 invented entities

The framework is empirical, so there are no formal derivations. The hidden costs are the free hyperparameters, the unverifiable data-quality premise, and the assumed reliability of self-generated pseudo-labels. No new physical entities are introduced; the Sparse Metric Prompt is an input construct whose external validity is not established.

free parameters (6)
  • Distance-balanced loss hyperparameter C = 400
    Eq. 5; selected in Sec. 5.7 using the DIODE benchmark (Tab. 15 right), and DIODE is later used as an evaluation set in Tabs. 5-6—tuning on a test set.
  • Teacher loss weights alpha, beta = alpha=15, beta=5
    Eq. 13; set without an independent validation protocol; affects the pretrained teacher.
  • Student loss weights gamma, delta = gamma=10, delta=2
    Eq. 15; set by hand and not independently justified.
  • Robust MAE top-n drop = 20%
    Eq. 8; discards the largest-error pixels per image. If errors are systematic rather than random, this can discard signal instead of noise.
  • Sparse prompt sample count N = 2,000-40,000
    Sec. 5.6; chosen for accuracy/efficiency trade-offs; affects all prompt-based results.
  • Data mix proportions = unreported
    Sec. 3.1 says 'a small portion' of rendered data but gives no per-source counts or balance; the composition is a free choice that affects the scaling trend.
axioms (5)
  • domain assumption All aggregated 3D sources can be converted to per-pixel metric depth via known calibration and have valid metric scale.
    Sec. 3.1: raw point clouds are projected using camera intrinsics and sensor-to-camera pose. If any large source has systematic scale/calibration error, the robust MAE loss can discard signal.
  • domain assumption The pre-trained depth model DepthPro [10] provides a reliable prior P_d for prompt preparation, and PDSA/GMDR preserve metric consistency.
    Sec. 3.2: prompt preparation is anchored to DepthPro predictions; the model therefore inherits DepthPro's errors and scale behavior.
  • domain assumption Teacher-generated pseudo-labels are accurate across near and far ranges and do not propagate teacher systematic errors.
    Sec. 3.3: the prompt-free student is trained entirely on teacher pseudo-labels, so its ceiling is the teacher's metric accuracy.
  • domain assumption Training and evaluation sets have no shared origin, as asserted for the zero-shot protocol.
    Sec. 7.1 states the only exception is ScanNet, but no overlap checks or dataset manifest are provided to verify this.
  • ad hoc to paper Random sparse masking provides a universal interface that decouples spatial reasoning from sensor and camera bias.
    Sec. 3.2: this is the core design hypothesis; it is supported only by in-paper empirical results, not by independent evidence.
invented entities (1)
  • Sparse Metric Prompt no independent evidence
    purpose: Universal sparse-conditioning interface intended to decouple spatial reasoning from sensor/camera bias.
    This is the paper's central invented construct. Its utility is evidenced only by in-paper ablations; no external falsifiable handle or released artifact is provided in the manuscript.

pith-pipeline@v1.3.0-alltime-deepseek · 34656 in / 18382 out tokens · 202018 ms · 2026-08-03T06:43:59.952216+00:00 · methodology

0 comments
read the original abstract

Scaling has powered recent advances in vision foundation models, yet extending this paradigm to metric depth estimation remains challenging due to heterogeneous sensor noise, camera-dependent biases, and metric ambiguity in noisy cross-source 3D data. We introduce Metric Anything, a simple and scalable pretraining framework that learns metric depth from noisy, diverse 3D sources without manually engineered prompts, camera-specific modeling, or task-specific architectures. Central to our approach is the Sparse Metric Prompt, created by randomly masking depth maps, which serves as a universal interface that decouples spatial reasoning from sensor and camera biases. Using about 20M image-depth pairs spanning reconstructed, captured, and rendered 3D data across 10000 camera models, we demonstrate-for the first time-a clear scaling trend in the metric depth track. The pretrained model excels at prompt-driven tasks such as depth completion, super-resolution and Radar-camera fusion, while its distilled prompt-free student achieves state-of-the-art results on monocular depth estimation, camera intrinsics recovery, single/multi-view metric 3D reconstruction, and VLA planning. We also show that using pretrained ViT of Metric Anything as a visual encoder significantly boosts Multimodal Large Language Model capabilities in spatial intelligence. These results show that metric depth estimation can benefit from the same scaling laws that drive modern foundation models, establishing a new path toward scalable and efficient real-world metric perception. We open-source MetricAnything at http://metric-anything.github.io/metric-anything-io/ to support community research.

Figures

Figures reproduced from arXiv: 2601.22054 by Baorui Ma, Donglin Di, Hao Li, Jiahui Yang, Jianxun Cui, Wei Chen, Xuancheng Zhang, Yan Xie.

Figure 1
Figure 1. Figure 1: Overview of Metric Anything. (I) We aggregate diverse open-source 3D data into per-pixel metric depth maps, forming a ∼20M image–depth dataset captured by over 10,000 cameras across heterogeneous scenes. (II) Sparse Metric Prompts, generated by randomly masking depth maps, provide a minimal interface that decouples spatial reasoning from sensor and camera biases, enabling metric depth learning from noisy, … view at source ↗
Figure 2
Figure 2. Figure 2: Scaling and Generalization. MetricAnything exhibits a clear scaling trend and strong overall down￾stream performance. • Demonstrated Scaling: Aggregation of 20M diverse depth-image pairs reveals stable scaling trends in metric depth estimation, previously unseen in this domain. • Universal Generalization: Both pretrained and distilled models achieve state-of-the-art performance across ten downstream tasks … view at source ↗
Figure 3
Figure 3. Figure 3: Percentile Depth Range Comparison from Seven Datasets (Real-world vs. Our Pseudo Labels). [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Skip-Connection in ViT-DPT Architecture. [PITH_FULL_IMAGE:figures/full_fig_p007_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Visualization of Depth SR and Completion. Our method better recovers missing regions with improved structure [PITH_FULL_IMAGE:figures/full_fig_p009_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Zero-shot Visual Comparisons on Challenging Test Samples. Our model robustly captures details of thin structures and in scenes with difficult lighting where competitors often fail. robust. This suggests that our model effectively handles large-scale depth variations typical of outdoor environments. • Robustness on Unconventional Data: On Booster, a dataset known for challenging lighting and tex￾tures, our … view at source ↗
Figure 7
Figure 7. Figure 7: Qualitative Comparison of Monocular Depth Estimation. Compared with MoGe2 and UniDepthv2, our distilled model produces more detailed and geometrically plausible predictions for both depth maps and point maps. Overall, while some baselines excel in specific niches, our method delivers the most balanced and consistently high performance across the full spectrum of test scenarios. Monocular metric point map. … view at source ↗
Figure 8
Figure 8. Figure 8: Qualitative Comparison of Point Maps. The red arrows indicate the GT distance, the yellow arrows indicate the distance from predicted point map. Monocular 3D geometry estimation aims to recover a per-pixel 3D point map in the camera coordinate system. In this setting, we leverage pseudo-labels generated by our pre-trained model to fine-tune recent state-of-the-art frameworks such as MoGe-2 [110], denoted a… view at source ↗
Figure 9
Figure 9. Figure 9: Qualitative Comparison of Depth Map. Compared with MoGe2 and UniDepthv2, our Student￾PointMap model, which is finetuned from MoGe2 using pseudo-labels predicted by the our pretrained model, achieves more stable and accurate depth estimation. 15 [PITH_FULL_IMAGE:figures/full_fig_p015_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: Auxiliary monocular depth inputs improve performance of MapAnything. The [PITH_FULL_IMAGE:figures/full_fig_p017_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: Enhancing VLA Planning with Metric Anything. We distill the depth-perception capability of Metric Anything into the VLA model by supervising it to predict metric-aware depth tokens. 17 [PITH_FULL_IMAGE:figures/full_fig_p017_11.png] view at source ↗
Figure 12
Figure 12. Figure 12: Enhancing 3D Spatial Reasoning with a Frozen ViT from Metric Anything . We evaluate our approach on the VIS Benchmark, covering video question-answering tasks like estimating object size, object’s distances, appearance order, route planning, and room size. Compared to mainstream large models, our method demonstrates robust and superior performance in 3D spatial understanding. 19 [PITH_FULL_IMAGE:figures/… view at source ↗
Figure 13
Figure 13. Figure 13: Enhancing 3D Spatial Reasoning in MLLMs. We enhance VLM capabilities by employing the frozen, pretrained ViT from Metric Anything as the visual encoder, thereby preserving its strong spatial under￾standing during fine-tuning. We report the evaluation results on VSI-Bench [123] in the Tab.11. For Spatial-MLLM and Qwen2.5VL-series [3], we use 16 frames as input. For other open-source methods and GPT-4o [42]… view at source ↗
Figure 14
Figure 14. Figure 14: Test-time Resolution Scaling. Qualitative results of depth estimation on an example image at 1×, 3×, and 9× the base input resolution. Higher resolutions recover finer details [PITH_FULL_IMAGE:figures/full_fig_p022_14.png] view at source ↗
Figure 15
Figure 15. Figure 15: Analysis of Our Proposed Distance-Balanced Loss. Left: The training loss for different C values. Right: The loss function curves. (Eq. 5 in the main manuscript ). Image: Night Env LiDAR Prompt Metric Depth [PITH_FULL_IMAGE:figures/full_fig_p023_15.png] view at source ↗
Figure 16
Figure 16. Figure 16: Robustness in Night-Time Driving. We deployed a test vehicle to evaluate performance under low￾light conditions. As scene brightness drops, visual signals deteriorate and object details fade. Despite this severe degradation, our model maintains remarkably robust. 5.7 Balance Weights We explore the effect of the balance weight C in Eq.14 in the main manuscript. We try different candidates including {50, 10… view at source ↗
Figure 17
Figure 17. Figure 17: Sensor Configuration for Real-World Generalization Evaluation. Our real-world test vehicle is equipped with three cameras (front, left-front, right-front) and a 128-beam solid-state LiDAR. Due to the LiDAR’s limited vertical field of view(pitch angle limitation), its captured point cloud does not fully cover the cameras’ combined frustums, leaving large image regions without metric depth cues. 6 Generaliz… view at source ↗
Figure 18
Figure 18. Figure 18 [PITH_FULL_IMAGE:figures/full_fig_p025_18.png] view at source ↗
Figure 19
Figure 19. Figure 19: Robustness in Adverse Weather. In the real-world deployment, we used a test vehicle to evaluate our pre-trained model for depth super-resolution and completion in rainy and foggy weather conditions without fine￾tuning. These adverse conditions significantly affect scene reflectance, causing the LiDAR to produce numerous artifacts or completely occlude critical objects. For example, the degraded data can l… view at source ↗
Figure 20
Figure 20. Figure 20: Generalization to Unseen Visual Domains. Depth prediction results on fisheye images, an unseen domain characterized by severe radial distortion. The model was applied in a zero-shot setting without fine-tuning [PITH_FULL_IMAGE:figures/full_fig_p027_20.png] view at source ↗
Figure 21
Figure 21. Figure 21: Generalization to Unseen Visual Domains. Depth prediction visualization for diverse in-the-wild images. 27 [PITH_FULL_IMAGE:figures/full_fig_p027_21.png] view at source ↗
Figure 22
Figure 22. Figure 22: Generalization to Unseen Visual Domains. Visualizing depth predictions on panoramic images, an unseen domain during training. Our model successfully handles such extreme distortion and novel viewpoints. 28 [PITH_FULL_IMAGE:figures/full_fig_p028_22.png] view at source ↗
Figure 23
Figure 23. Figure 23: Generalization to Unseen Visual Domains. Additional visualizations of depth predictions on diverse in-the-wild images. 29 [PITH_FULL_IMAGE:figures/full_fig_p029_23.png] view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 5 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. DrivingDepth: Sparse-Prompted Pixel-wise Scale Correction for Driving Depth Estimation

    cs.CV 2026-06 unverdicted novelty 7.0

    DrivingDepth achieves SOTA metric depth on nuScenes by residual pixel-wise scale correction on frozen foundation models using sparse LiDAR prompts, preserving geometric consistency.

  2. Any to Full: Prompting Depth Anything for Depth Completion in One Stage

    cs.CV 2026-03 accept novelty 6.5

    Any2Full reformulates depth completion as one-stage scale-prompting of a pretrained monocular depth estimator, yielding domain-general, pattern-agnostic dense metric depth with lower error and higher speed than prior methods.

  3. SUFLECA: Scaling Up Feature Learning for CAD-to-image Alignment

    cs.CV 2026-07 conditional novelty 6.0

    SUFLECA achieves state-of-the-art single-image CAD-to-image alignment by scaling NOC-supervised feature learning to 674K images and adding geometrically consistent correspondence filtering.

  4. X-Lens: Real-Time Metric Depth Estimation with Heterogeneous Cameras

    cs.CV 2026-07 unverdicted novelty 6.0

    A 0.04B-parameter feed-forward model estimates metric depth from variable calibrated fisheye and pinhole views using calibration tokens and Jacobian distortion bias, with a new multi-view synthetic dataset.

  5. X-Lens: Real-Time Metric Depth Estimation with Heterogeneous Cameras

    cs.CV 2026-07 conditional novelty 6.0

    X-Lens fuses arbitrary calibrated fisheye and pinhole views into real-time metric depth at 41 FPS with a 0.04B-parameter model and a new 266K-frame synthetic dataset.

Reference graph

Works this paper leans on

139 extracted references · 32 linked inside Pith · cited by 4 Pith papers

  1. [1]

    Mapillary planet-scale depth dataset

    Manuel López Antequera, Pau Gargallo, Markus Hofinger, Samuel Rota Bulo, Yubin Kuang, and Peter Kontschieder. Mapillary planet-scale depth dataset. InEuropean Conference on Computer Vision, pages 589–604. Springer, 2020

    2020

  2. [2]

    Apollo synthetic dataset, 2019

    Baidu Apollo. Apollo synthetic dataset, 2019. Accessed: 2025-6-12

    2019

  3. [3]

    Qwen2.5-vl technical report.ArXiv, abs/2502.13923, 2025

    Shuai Bai, Keqin Chen, Xuejing Liu, Jialin Wang, Wenbin Ge, Sibo Song, Kai Dang, Peng Wang, Shijie Wang, Jun Tang, Humen Zhong, Yuanzhi Zhu, Mingkun Yang, Zhaohai Li, Jianqiang Wan, Pengfei Wang, Wei Ding, Zheren Fu, Yiheng Xu, Jiabo Ye, Xi Zhang, Tianbao Xie, Zesen Cheng, Hang Zhang, Zhibo Yang, Haiyang Xu, and Junyang Lin. Qwen2.5-vl technical report.Ar...

    Pith/arXiv arXiv 2025

  4. [4]

    Arkitscenes: A diverse real-world dataset for 3d indoor scene understanding using mobile rgb-d data.arXiv preprint arXiv:2111.08897, 2021

    Gilad Baruch, Zhuoyuan Chen, Afshin Dehghan, Tal Dimry, Yuri Feigin, Peter Fu, Thomas Gebauer, Bran- don Joffe, Daniel Kurz, Arik Schwartz, et al. Arkitscenes: A diverse real-world dataset for 3d indoor scene understanding using mobile rgb-d data.arXiv preprint arXiv:2111.08897, 2021

    Pith/arXiv arXiv 2021

  5. [5]

    Uasol, a large-scale high-resolution outdoor stereo dataset.Scientific data, 6(1):162, 2019

    Zuria Bauer, Francisco Gomez-Donoso, Edmanuel Cruz, Sergio Orts-Escolano, and Miguel Cazorla. Uasol, a large-scale high-resolution outdoor stereo dataset.Scientific data, 6(1):162, 2019

    2019

  6. [6]

    Adabins: Depth estimation using adaptive bins

    Shariq Farooq Bhat, Ibraheem Alhashim, and Peter Wonka. Adabins: Depth estimation using adaptive bins. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 4009–4018, 2021

    2021

  7. [7]

    Localbins: Improving depth estimation by learning local distributions

    Shariq Farooq Bhat, Ibraheem Alhashim, and Peter Wonka. Localbins: Improving depth estimation by learning local distributions. InEuropean Conference on Computer Vision, pages 480–496. Springer, 2022

    2022

  8. [8]

    Zoedepth: Zero-shot transfer by combining relative and metric depth, 2023

    Shariq Farooq Bhat, Reiner Birkl, Diana Wofk, Peter Wonka, and Matthias Müller. Zoedepth: Zero-shot transfer by combining relative and metric depth, 2023

    2023

  9. [9]

    3d cavla: Leveraging depth and 3d context to generalize vision language action models for unseen tasks.arXiv preprint arXiv:2505.05800, 2025

    Vineet Bhat, Yu-Hsiang Lan, Prashanth Krishnamurthy, Ramesh Karri, and Farshad Khorrami. 3d cavla: Leveraging depth and 3d context to generalize vision language action models for unseen tasks.arXiv preprint arXiv:2505.05800, 2025

    arXiv 2025

  10. [10]

    Richter, and Vladlen Koltun

    Aleksei Bochkovskii, Amaël Delaunoy, Hugo Germain, Marcel Santos, Yichao Zhou, Stephan R. Richter, and Vladlen Koltun. Depth pro: Sharp monocular metric depth in less than a second. InInternational Conference on Learning Representations, 2025. 33

    2025

  11. [11]

    A naturalistic open source movie for optical flow evaluation

    Daniel J Butler, Jonas Wulff, Garrett B Stanley, and Michael J Black. A naturalistic open source movie for optical flow evaluation. InEuropean conference on computer vision, pages 611–625. Springer, 2012

    2012

  12. [12]

    nuscenes: A multimodal dataset for autonomous driving

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

    2020

  13. [13]

    Worldvla: Towards autoregressive action world model.arXiv preprint arXiv:2506.21539, 2025

    Jun Cen, Chaohui Yu, Hangjie Yuan, Yuming Jiang, Siteng Huang, Jiayan Guo, Xin Li, Yibing Song, Hao Luo, Fan Wang, et al. Worldvla: Towards autoregressive action world model.arXiv preprint arXiv:2506.21539, 2025

    Pith/arXiv arXiv 2025

  14. [14]

    Matterport3d: Learning from rgb-d data in indoor environments

    Angel Chang, Angela Dai, Thomas Funkhouser, Maciej Halber, Matthias Niessner, Manolis Savva, Shuran Song, Andy Zeng, and Yinda Zhang. Matterport3d: Learning from rgb-d data in indoor environments. International Conference on 3D Vision (3DV), 2017

    2017

  15. [15]

    Single-image depth perception in the wild.Advances in neural information processing systems, 29, 2016

    Weifeng Chen, Zhao Fu, Dawei Yang, and Jia Deng. Single-image depth perception in the wild.Advances in neural information processing systems, 29, 2016

    2016

  16. [16]

    Oasis: A large- scale dataset for single image 3d in the wild

    Weifeng Chen, Shengyi Qian, David Fan, Noriyuki Kojima, Max Hamilton, and Jia Deng. Oasis: A large- scale dataset for single image 3d in the wild. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), June 2020

    2020

  17. [17]

    Internvl: Scaling up vision foundation models and aligning for generic visual- linguistic tasks

    Zhe Chen, Jiannan Wu, Wenhai Wang, Weijie Su, Guo Chen, Sen Xing, Muyan Zhong, Qinglong Zhang, Xizhou Zhu, Lewei Lu, et al. Internvl: Scaling up vision foundation models and aligning for generic visual- linguistic tasks. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 24185–24198, 2024

    2024

  18. [18]

    Cspn++: Learning context and resource aware convolutional spatial propagation networks for depth completion

    Xinjing Cheng, Peng Wang, Chenye Guan, and Ruigang Yang. Cspn++: Learning context and resource aware convolutional spatial propagation networks for depth completion. InAAAI, 2020

    2020

  19. [19]

    Learning depth with convolutional spatial propagation network.TPAMI, 2019

    Xinjing Cheng, Peng Wang, and Ruigang Yang. Learning depth with convolutional spatial propagation network.TPAMI, 2019

    2019

  20. [20]

    The cityscapes dataset for semantic urban scene understanding

    Marius Cordts, Mohamed Omran, Sebastian Ramos, Timo Rehfeld, Markus Enzweiler, Rodrigo Benenson, Uwe Franke, Stefan Roth, and Bernt Schiele. The cityscapes dataset for semantic urban scene understanding. InProceedings of the IEEE conference on computer vision and pattern recognition, pages 3213–3223, 2016

    2016

  21. [21]

    Scannet: Richly-annotated 3d reconstructions of indoor scenes

    Angela Dai, Angel X Chang, Manolis Savva, Maciej Halber, Thomas Funkhouser, and Matthias Nießner. Scannet: Richly-annotated 3d reconstructions of indoor scenes. InProceedings of the IEEE conference on computer vision and pattern recognition, pages 5828–5839, 2017

    2017

  22. [22]

    Fu, Stefano Ermon, Atri Rudra, and Christopher Ré

    Tri Dao, Daniel Y . Fu, Stefano Ermon, Atri Rudra, and Christopher Ré. FlashAttention: Fast and memory-efficient exact attention with IO-awareness. InAdvances in Neural Information Processing Systems (NeurIPS), 2022

    2022

  23. [23]

    An image is worth 16x16 words: Transformers for image recognition at scale.ICLR, 2021

    Alexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn, Xiaohua Zhai, Thomas Un- terthiner, Mostafa Dehghani, Matthias Minderer, Georg Heigold, Sylvain Gelly, Jakob Uszkoreit, and Neil Houlsby. An image is worth 16x16 words: Transformers for image recognition at scale.ICLR, 2021

    2021

  24. [24]

    Predicting depth, surface normals and semantic labels with a common multi- scale convolutional architecture

    David Eigen and Rob Fergus. Predicting depth, surface normals and semantic labels with a common multi- scale convolutional architecture. InProceedings of the IEEE international conference on computer vision, pages 2650–2658, 2015

    2015

  25. [25]

    Depth map prediction from a single image using a multi- scale deep network.Advances in neural information processing systems, 27, 2014

    David Eigen, Christian Puhrsch, and Rob Fergus. Depth map prediction from a single image using a multi- scale deep network.Advances in neural information processing systems, 27, 2014

    2014

  26. [26]

    Mid-air: A multi-modal dataset for extremely low altitude drone flights

    Michael Fonder and Marc Van Droogenbroeck. Mid-air: A multi-modal dataset for extremely low altitude drone flights. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition workshops, pages 0–0, 2019

    2019

  27. [27]

    Deep ordinal regres- sion network for monocular depth estimation

    Huan Fu, Mingming Gong, Chaohui Wang, Kayhan Batmanghelich, and Dacheng Tao. Deep ordinal regres- sion network for monocular depth estimation. InProceedings of the IEEE conference on computer vision and pattern recognition, pages 2002–2011, 2018

    2002

  28. [28]

    Virtual worlds as proxy for multi-object tracking analysis

    Adrien Gaidon, Qiao Wang, Yohann Cabon, and Eleonora Vig. Virtual worlds as proxy for multi-object tracking analysis. InProceedings of the IEEE conference on computer vision and pattern recognition, pages 4340–4349, 2016

    2016

  29. [29]

    R4dyn: Exploring radar for self-supervised monocular depth estimation of dynamic scenes

    Stefano Gasperini, Patrick Koch, Vinzenz Dallabetta, Nassir Navab, Benjamin Busam, and Federico Tombari. R4dyn: Exploring radar for self-supervised monocular depth estimation of dynamic scenes. In 2021 International Conference on 3D Vision (3DV), pages 751–760. IEEE, 2021. 34

    2021

  30. [30]

    Dsec: A stereo event camera dataset for driving scenarios.IEEE Robotics and Automation Letters, 6(3):4947–4954, 2021

    Mathias Gehrig, Willem Aarents, Daniel Gehrig, and Davide Scaramuzza. Dsec: A stereo event camera dataset for driving scenarios.IEEE Robotics and Automation Letters, 6(3):4947–4954, 2021

    2021

  31. [31]

    Are we ready for autonomous driving.The KITTI vision benchmark suite

    A Geiger, P Lenz, and R Urtasun. Are we ready for autonomous driving.The KITTI vision benchmark suite. InCVPR, 2:5, 2012

    2012

  32. [32]

    Digging into self-supervised monocular depth estimation

    Clément Godard, Oisin Mac Aodha, Michael Firman, and Gabriel J Brostow. Digging into self-supervised monocular depth estimation. InProceedings of the IEEE/CVF international conference on computer vision, pages 3828–3838, 2019

    2019

  33. [33]

    All for one, and one for all: Urbansyn dataset, the third musketeer of synthetic driving scenes.Neurocomputing, 637:130038, 2025

    Jose L Gómez, Manuel Silva, Antonio Seoane, Agnès Borrás, Mario Noriega, Germán Ros, Jose A Iglesias- Guitian, and Antonio M López. All for one, and one for all: Urbansyn dataset, the third musketeer of synthetic driving scenes.Neurocomputing, 637:130038, 2025

    2025

  34. [34]

    3d packing for self- supervised monocular depth estimation

    Vitor Guizilini, Rares Ambrus, Sudeep Pillai, Allan Raventos, and Adrien Gaidon. 3d packing for self- supervised monocular depth estimation. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 2485–2494, 2020

    2020

  35. [36]

    Towards zero-shot scale- aware monocular depth estimation

    Vitor Guizilini, Igor Vasiljevic, Dian Chen, Rares , Ambrus, , and Adrien Gaidon. Towards zero-shot scale- aware monocular depth estimation. InProceedings of the IEEE/CVF International Conference on Computer Vision, pages 9233–9243, 2023

    2023

  36. [37]

    One thousand and one hours: Self-driving motion prediction dataset.arXiv preprint arXiv:2006.14480, 2020

    John Houston, Guido Zuidhof, Luca Bergamini, Yawei Ye, Long Chen, Ashesh Jain, Sammy Omari, Vladimir Iglovikov, and Peter Ondruska. One thousand and one hours: Self-driving motion prediction dataset.arXiv preprint arXiv:2006.14480, 2020

    Pith/arXiv arXiv 2006

  37. [38]

    Mu Hu, Wei Yin, Chi Zhang, Zhipeng Cai, Xiaoxiao Long, Hao Chen, Kaixuan Wang, Gang Yu, Chunhua Shen, and Shaojie Shen. Metric3d v2: A versatile monocular geometric foundation model for zero-shot met- ric depth and surface normal estimation.IEEE Transactions on Pattern Analysis and Machine Intelligence, 2024

    2024

  38. [39]

    Thinkact: Vision-language-action reasoning via reinforced visual latent planning.arXiv preprint arXiv:2507.16815, 2025

    Chi-Pin Huang, Yueh-Hua Wu, Min-Hung Chen, Yu-Chiang Frank Wang, and Fu-En Yang. Thinkact: Vision-language-action reasoning via reinforced visual latent planning.arXiv preprint arXiv:2507.16815, 2025

    Pith/arXiv arXiv 2025

  39. [40]

    Deepmvs: Learning multi-view stereopsis

    Po-Han Huang, Kevin Matzen, Johannes Kopf, Narendra Ahuja, and Jia-Bin Huang. Deepmvs: Learning multi-view stereopsis. InProceedings of the IEEE conference on computer vision and pattern recognition, pages 2821–2830, 2018

    2018

  40. [41]

    Nora: A small open-sourced generalist vision language action model for embodied tasks.arXiv preprint arXiv:2504.19854, 2025

    Chia-Yu Hung, Qi Sun, Pengfei Hong, Amir Zadeh, Chuan Li, U Tan, Navonil Majumder, Soujanya Poria, et al. Nora: A small open-sourced generalist vision language action model for embodied tasks.arXiv preprint arXiv:2504.19854, 2025

    Pith/arXiv arXiv 2025

  41. [42]

    Gpt-4o system card.arXiv preprint arXiv:2410.21276, 2024

    Aaron Hurst, Adam Lerer, Adam P Goucher, Adam Perelman, Aditya Ramesh, Aidan Clark, AJ Ostrow, Akila Welihinda, Alan Hayes, Alec Radford, et al. Gpt-4o system card.arXiv preprint arXiv:2410.21276, 2024

    Pith/arXiv arXiv 2024

  42. [43]

    Perspective fields for single image camera calibration

    Linyi Jin, Jianming Zhang, Yannick Hold-Geoffroy, Oliver Wang, Kevin Blackburn-Matzen, Matthew Sticha, and David F Fouhey. Perspective fields for single image camera calibration. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 17307–17316, 2023

    2023

  43. [44]

    Repurposing diffusion-based image generators for monocular depth estimation

    Bingxin Ke, Anton Obukhov, Shengyu Huang, Nando Metzger, Rodrigo Caye Daudt, and Konrad Schindler. Repurposing diffusion-based image generators for monocular depth estimation. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 9492–9502, 2024

    2024

  44. [45]

    MapA- nything: Universal feed-forward metric 3D reconstruction, 2025

    Nikhil Keetha, Norman Müller, Johannes Schönberger, Lorenzo Porzi, Yuchen Zhang, Tobias Fischer, Arno Knapitsch, Duncan Zauss, Ethan Weber, Nelson Antunes, Jonathon Luiten, Manuel Lopez-Antequera, Samuel Rota Bulò, Christian Richardt, Deva Ramanan, Sebastian Scherer, and Peter Kontschieder. MapA- nything: Universal feed-forward metric 3D reconstruction, 2...

    Pith/arXiv arXiv 2025

  45. [46]

    Openvla: An open-source vision-language-action model.arXiv preprint arXiv:2406.09246, 2024

    Moo Jin Kim, Karl Pertsch, Siddharth Karamcheti, Ted Xiao, Ashwin Balakrishna, Suraj Nair, Rafael Rafailov, Ethan Foster, Grace Lam, Pannag Sanketi, et al. Openvla: An open-source vision-language-action model.arXiv preprint arXiv:2406.09246, 2024

    Pith/arXiv arXiv 2024

  46. [47]

    Segment anything

    Alexander Kirillov, Eric Mintun, Nikhila Ravi, Hanzi Mao, Chloe Rolland, Laura Gustafson, Tete Xiao, Spencer Whitehead, Alexander C Berg, Wan-Yen Lo, et al. Segment anything. InProceedings of the IEEE/CVF international conference on computer vision, pages 4015–4026, 2023. 35

    2023

  47. [48]

    Evaluation of cnn-based single- image depth estimation methods

    Tobias Koch, Lukas Liebel, Friedrich Fraundorfer, and Marco Korner. Evaluation of cnn-based single- image depth estimation methods. InProceedings of the European Conference on Computer Vision (ECCV) Workshops, pages 0–0, 2018

    2018

  48. [49]

    Pulling things out of perspective

    Lubor Ladicky, Jianbo Shi, and Marc Pollefeys. Pulling things out of perspective. InProceedings of the IEEE conference on computer vision and pattern recognition, pages 89–96, 2014

    2014

  49. [50]

    Deeper depth prediction with fully convolutional residual networks

    Iro Laina, Christian Rupprecht, Vasileios Belagiannis, Federico Tombari, and Nassir Navab. Deeper depth prediction with fully convolutional residual networks. In2016 Fourth international conference on 3D vision (3DV), pages 239–248. IEEE, 2016

    2016

  50. [51]

    Molmoact: Action reasoning models that can reason in space.arXiv preprint arXiv:2508.07917, 2025

    Jason Lee, Jiafei Duan, Haoquan Fang, Yuquan Deng, Shuo Liu, Boyang Li, Bohan Fang, Jieyu Zhang, Yi Ru Wang, Sangho Lee, et al. Molmoact: Action reasoning models that can reason in space.arXiv preprint arXiv:2508.07917, 2025

    Pith/arXiv arXiv 2025

  51. [52]

    Grounding image matching in 3d with mast3r

    Vincent Leroy, Yohann Cabon, and Jérôme Revaud. Grounding image matching in 3d with mast3r. In European Conference on Computer Vision, pages 71–91. Springer, 2024

    2024

  52. [53]

    Llava-onevision: Easy visual task transfer.arXiv preprint arXiv:2408.03326, 2024

    Bo Li, Yuanhan Zhang, Dong Guo, Renrui Zhang, Feng Li, Hao Zhang, Kaichen Zhang, Yanwei Li, Ziwei Liu, and Chunyuan Li. Llava-onevision: Easy visual task transfer.arXiv preprint arXiv:2408.03326, 2024

    Pith/arXiv arXiv 2024

  53. [54]

    Radarcam-depth: Radar-camera fusion for depth estimation with learned metric scale

    Han Li, Yukai Ma, Yaqing Gu, Kewei Hu, Yong Liu, and Xingxing Zuo. Radarcam-depth: Radar-camera fusion for depth estimation with learned metric scale. In2024 IEEE International Conference on Robotics and Automation (ICRA), pages 10665–10672. IEEE, 2024

    2024

  54. [55]

    Sparse beats dense: Rethinking supervision in radar-camera depth completion

    Huadong Li, Minhao Jing, Wang Jin, Shichao Dong, Jiajun Liang, Haoqiang Fan, and Renhe Ji. Sparse beats dense: Rethinking supervision in radar-camera depth completion. InEuropean Conference on Computer Vision, pages 127–143. Springer, 2024

    2024

  55. [56]

    Matrixcity: A large-scale city dataset for city-scale neural rendering and beyond

    Yixuan Li, Lihan Jiang, Linning Xu, Yuanbo Xiangli, Zhenzhi Wang, Dahua Lin, and Bo Dai. Matrixcity: A large-scale city dataset for city-scale neural rendering and beyond. InProceedings of the IEEE/CVF International Conference on Computer Vision, pages 3205–3215, 2023

    2023

  56. [57]

    Megadepth: Learning single-view depth prediction from internet photos

    Zhengqi Li and Noah Snavely. Megadepth: Learning single-view depth prediction from internet photos. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 2041–2050, 2018

    2041

  57. [58]

    Patchfusion: An end-to-end tile-based framework for high-resolution monocular metric depth estimation

    Zhenyu Li, Shariq Farooq Bhat, and Peter Wonka. Patchfusion: An end-to-end tile-based framework for high-resolution monocular metric depth estimation. InProceedings of the IEEE/CVF Conference on Com- puter Vision and Pattern Recognition, pages 10016–10025, 2024

    2024

  58. [59]

    Chen, Zhenyu Li, Guang Shi, Jiashi Feng, and Bingyi Kang

    Haotong Lin, Sili Chen, Jun Hao Liew, Donny Y . Chen, Zhenyu Li, Guang Shi, Jiashi Feng, and Bingyi Kang. Depth anything 3: Recovering the visual space from any views.arXiv preprint arXiv:2511.10647, 2025

    Pith/arXiv arXiv 2025

  59. [60]

    Prompting depth anything for 4k resolution accurate metric depth estimation

    Haotong Lin, Sida Peng, Jingxiao Chen, Songyou Peng, Jiaming Sun, Minghuan Liu, Hujun Bao, Jiashi Feng, Xiaowei Zhou, and Bingyi Kang. Prompting depth anything for 4k resolution accurate metric depth estimation. InCVPR, 2025

    2025

  60. [61]

    Vila: On pre-training for visual language models.2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 26679–26689, 2023

    Ji Lin, Hongxu Yin, Wei Ping, Yao Lu, Pavlo Molchanov, Andrew Tao, Huizi Mao, Jan Kautz, Mohammad Shoeybi, and Song Han. Vila: On pre-training for visual language models.2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 26679–26689, 2023

    2024

  61. [62]

    Depth estimation from monocular images and sparse radar data

    Juan-Ting Lin, Dengxin Dai, and Luc Van Gool. Depth estimation from monocular images and sparse radar data. in 2020 ieee. InRSJ International Conference on Intelligent Robots and Systems (IROS), pages 10233–10240, 2020

    2020

  62. [63]

    Libero: Bench- marking knowledge transfer for lifelong robot learning.Advances in Neural Information Processing Sys- tems, 36:44776–44791, 2023

    Bo Liu, Yifeng Zhu, Chongkai Gao, Yihao Feng, Qiang Liu, Yuke Zhu, and Peter Stone. Libero: Bench- marking knowledge transfer for lifelong robot learning.Advances in Neural Information Processing Sys- tems, 36:44776–44791, 2023

    2023

  63. [64]

    Depthlab: From partial to complete.arXiv preprint arXiv:2412.18153, 2024

    Zhiheng Liu, Ka Leong Cheng, Qiuyu Wang, Shuzhe Wang, Hao Ouyang, Bin Tan, Kai Zhu, Yujun Shen, Qifeng Chen, and Ping Luo. Depthlab: From partial to complete.arXiv preprint arXiv:2412.18153, 2024

    arXiv 2024

  64. [65]

    Depth estimation from monocular images and sparse radar using deep ordinal regression network

    Chen-Chou Lo and Patrick Vandewalle. Depth estimation from monocular images and sparse radar using deep ordinal regression network. In2021 IEEE International Conference on Image Processing (ICIP), pages 3343–3347. IEEE, 2021

    2021

  65. [66]

    Rcdpt: Radar-camera fusion dense prediction transformer

    Chen-Chou Lo and Patrick Vandewalle. Rcdpt: Radar-camera fusion dense prediction transformer. In ICASSP 2023-2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pages 1–5. IEEE, 2023. 36

    2023

  66. [67]

    Radar-camera pixel depth association for depth completion

    Yunfei Long, Daniel Morris, Xiaoming Liu, Marcos Castro, Punarjay Chakravarty, and Praveen Narayanan. Radar-camera pixel depth association for depth completion. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 12507–12516, 2021

    2021

  67. [68]

    You see it, you got it: Learning 3d creation on pose-free videos at scale

    Baorui Ma, Huachen Gao, Haoge Deng, Zhengxiong Luo, Tiejun Huang, Lulu Tang, and Xinlong Wang. You see it, you got it: Learning 3d creation on pose-free videos at scale. InProceedings of the Computer Vision and Pattern Recognition Conference, pages 2016–2029, 2025

    2016

  68. [69]

    Spring: A high- resolution high-detail dataset and benchmark for scene flow, optical flow and stereo

    Lukas Mehl, Jenny Schmalfuss, Azin Jahedi, Yaroslava Nalivayko, and Andrés Bruhn. Spring: A high- resolution high-detail dataset and benchmark for scene flow, optical flow and stereo. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 4981–4991, 2023

    2023

  69. [70]

    Dinov2: Learning robust visual features without supervision.arXiv preprint arXiv:2304.07193, 2023

    Maxime Oquab, Timothée Darcet, Théo Moutakanni, Huy V o, Marc Szafraniec, Vasil Khalidov, Pierre Fernandez, Daniel Haziza, Francisco Massa, Alaaeldin El-Nouby, et al. Dinov2: Learning robust visual features without supervision.arXiv preprint arXiv:2304.07193, 2023

    Pith/arXiv arXiv 2023

  70. [71]

    Depth prompting for sensor-agnostic depth estimation

    Jin-Hwi Park, Chanhwi Jeong, Junoh Lee, and Hae-Gon Jeon. Depth prompting for sensor-agnostic depth estimation. InCVPR, 2024

    2024

  71. [72]

    Scalable diffusion models with transformers

    William Peebles and Saining Xie. Scalable diffusion models with transformers. InProceedings of the IEEE/CVF international conference on computer vision, pages 4195–4205, 2023

    2023

  72. [73]

    Fast: Efficient action tokenization for vision-language-action models.arXiv preprint arXiv:2501.09747, 2025

    Karl Pertsch, Kyle Stachowicz, Brian Ichter, Danny Driess, Suraj Nair, Quan Vuong, Oier Mees, Chelsea Finn, and Sergey Levine. Fast: Efficient action tokenization for vision-language-action models.arXiv preprint arXiv:2501.09747, 2025

    Pith/arXiv arXiv 2025

  73. [74]

    Unidepthv2: Universal monocular metric depth estimation made simpler.arXiv preprint arXiv:2502.20110, 2025

    Luigi Piccinelli, Christos Sakaridis, Yung-Hsu Yang, Mattia Segu, Siyuan Li, Wim Abbeloos, and Luc Van Gool. Unidepthv2: Universal monocular metric depth estimation made simpler.arXiv preprint arXiv:2502.20110, 2025

    Pith/arXiv arXiv 2025

  74. [75]

    Unidepth: Universal monocular metric depth estimation

    Luigi Piccinelli, Yung-Hsu Yang, Christos Sakaridis, Mattia Segu, Siyuan Li, Luc Van Gool, and Fisher Yu. Unidepth: Universal monocular metric depth estimation. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 10106–10116, 2024

    2024

  75. [76]

    Spatialvla: Exploring spatial representations for visual-language-action model

    Delin Qu, Haoming Song, Qizhi Chen, Yuanqi Yao, Xinyi Ye, Yan Ding, Zhigang Wang, JiaYuan Gu, Bin Zhao, Dong Wang, et al. Spatialvla: Exploring spatial representations for visual-language-action model. arXiv preprint arXiv:2501.15830, 2025

    Pith/arXiv arXiv 2025

  76. [77]

    Learning transferable visual models from natural language supervision

    Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, et al. Learning transferable visual models from natural language supervision. InInternational conference on machine learning, pages 8748–8763. PmLR, 2021

    2021

  77. [78]

    Habitat-matterport 3d dataset (hm3d): 1000 large-scale 3d environments for embodied ai.arXiv preprint arXiv:2109.08238, 2021

    Santhosh K Ramakrishnan, Aaron Gokaslan, Erik Wijmans, Oleksandr Maksymets, Alex Clegg, John Turner, Eric Undersander, Wojciech Galuba, Andrew Westbury, Angel X Chang, et al. Habitat-matterport 3d dataset (hm3d): 1000 large-scale 3d environments for embodied ai.arXiv preprint arXiv:2109.08238, 2021

    Pith/arXiv arXiv 2021

  78. [79]

    Zero-shot text-to-image generation

    Aditya Ramesh, Mikhail Pavlov, Gabriel Goh, Scott Gray, Chelsea V oss, Alec Radford, Mark Chen, and Ilya Sutskever. Zero-shot text-to-image generation. InInternational conference on machine learning, pages 8821–8831. Pmlr, 2021

    2021

  79. [80]

    Booster: a benchmark for depth from images of specular and transparent surfaces.IEEE Transactions on Pattern Analysis and Machine Intelligence, 46(1):85–102, 2023

    Pierluigi Zama Ramirez, Alex Costanzino, Fabio Tosi, Matteo Poggi, Samuele Salti, Stefano Mattoccia, and Luigi Di Stefano. Booster: a benchmark for depth from images of specular and transparent surfaces.IEEE Transactions on Pattern Analysis and Machine Intelligence, 46(1):85–102, 2023

    2023

  80. [81]

    Vision transformers for dense prediction

    René Ranftl, Alexey Bochkovskiy, and Vladlen Koltun. Vision transformers for dense prediction. InPro- ceedings of the IEEE/CVF international conference on computer vision, pages 12179–12188, 2021

    2021

Showing first 80 references.