Pith. sign in

REVIEW 4 major objections 4 minor 1 cited by

PatchRefiner V2: Fast and Lightweight Real-Domain High-Resolution Metric Depth Estimation

T0 review · 4 major / 4 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read A lightweight refiner guided by coarse depth features can beat heavy tile-based refinement on 4K depth accuracy while cutting inference time.

desk verdict Solid engineering follow-up to PatchRefiner, but Eq. (6) is identically zero as written and the SOTA claim is overstated; worth reviewing after fixes. read the letter →

arxiv 2501.01121 v1 pith:WTWOEV7S submitted 2025-01-02 cs.CV

Zhenyu Li , Wenqing Cui , Shariq Farooq Bhat , Peter Wonka This is my paper · ORCID
classification cs.CV
keywords monoculardepthestimationhigh-resolutiontile-basedrefinementcoarse-to-finemoduleGuidedDenoisingUnitNoisyPretrainingscale-and-shiftinvariantgradientmatchingsynthetic-to-realtransfer
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

PatchRefiner V2 asks whether the expensive second pass in tile-based high-resolution depth estimation is necessary. The paper's answer is that a lightweight encoder can replace the heavy depth foundation model in the refiner branch, provided the lost depth-aligned features are restored by two new mechanisms: a coarse-to-fine module that uses global coarse-depth features to denoise local patch features, and a pretraining stage that feeds the refiner random noise instead of coarse features. On the synthetic 4K benchmark UnrealStereo4K, the largest variant PRV2C achieves an RMSE of 0.884, a new state-of-the-art, while running 2.3x faster than the previous PatchRefiner; the smallest variant is 9.2x smaller and 10.7x faster than PatchFusion. A third contribution, a scale-and-shift invariant gradient-matching loss for synthetic-to-real transfer, improves boundary F1 scores on Cityscapes by 17.2% for PatchRefiner and 20.9% for PRV2E. If these results hold, high-resolution metric depth estimation no longer needs to trade accuracy for speed.

What carries the argument

The central mechanism is the Coarse-to-Fine (C2F) module, a bottom-up stack of layers each containing a Guided Denoising Unit (GDU) and a residual convolutional unit. The GDU concatenates coarse features with fine refiner features, passes them through a convolutional block and sigmoid to produce a 0-to-1 weight map, and multiplies that map into the fine features, denoising them before the Fine-to-Coarse (F2C) module injects fine detail into the coarse depth map. Around this sit two supporting mechanisms: Noisy Pretraining, which feeds random normal noise as coarse-guidance input so the refiner branch learns to extract depth-relevant features on its own, and the Scale-and-Shift Invariant Gradient Matching (SSIGM) loss, which aligns predicted and pseudo-label depths by least-squares scale/shift and then penalizes gradient differences so high-frequency boundaries are learned directly.

What would settle it

Run PRV2 on a real-domain collection where the coarse base model produces systematically biased depth (for example, night-driving or rain images outside its training distribution) and test whether the refiner removes the bias; the paper's premise predicts it will not, because the refiner only adds local residuals to the coarse depth map.

Watch

Extended reading notes

Core claim

On the paper's own terms, the discovery is that the refinement stage of tile-based depth estimation can be made fast and small without sacrificing accuracy, because what the lightweight refiner lacks is not capacity but an aligned feature representation. The C2F module supplies that alignment by letting coarse features act as a gating signal: a Guided Denoising Unit computes a sigmoid weight map from concatenated coarse and fine features and multiplies it into the fine shortcut features, suppressing noise while preserving high-frequency structure. Noisy Pretraining then pretrains the whole refiner branch with random coarse features, forcing it to learn depth-relevant features without relying on the coarse branch. Finally, SSIGM replaces the scale-and-shift invariant MAE with a gradient-matching term after least-squares scale/shift alignment, so pseudo-label supervision directly targets high-frequency detail. The combination yields state-of-the-art RMSE 0.884 on UnrealStereo4K, larger boundary F1 gains on Cityscapes, and consistent improvements when the base model is either ZoeDepth or Depth Anything V2.

Load-bearing premise

The coarse branch already provides a reliable base depth estimate, so the refiner only needs to add high-frequency detail; if the coarse depth is badly wrong in a target domain, the lightweight refiner cannot repair it.

Editorial extensions

If this is right

  • PRV2C sets a new state-of-the-art RMSE of 0.884 on UnrealStereo4K while running 2.3x faster than PatchRefiner.
  • PRV2M shows that a 47.0M-parameter refiner can improve RMSE by 22.2% over the coarse baseline while being 9.2x smaller and 10.7x faster than PatchFusion.
  • The SSIGM loss improves boundary F1 by 17.2% on PatchRefiner and 20.9% on PRV2E on Cityscapes without hurting scale accuracy.
  • The framework is base-model agnostic: it improves both ZoeDepth and Depth Anything V2 coarse branches on UnrealStereo4K.
  • End-to-end training of the full framework becomes feasible, replacing the stage-wise training required by earlier tile-based methods.

Reading between the lines

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

  • The Noisy Pretraining recipe may transfer to other dual-branch dense prediction tasks where a global branch guides local processing; its only requirement is that guidance features can be replaced by random noise during pretraining.
  • The SSIGM design suggests a general ordering principle for gradient-based domain transfer: align global scale and shift first, then match gradients; the paper's ablation shows the reverse order (GMSSI) is worse on boundary metrics.
  • A testable implication of the coarse-branch assumption is that PRV2's advantage should shrink when the coarse base model is unreliable in the target domain, since the refiner only adds residuals to $D_c$ and cannot repair global bias.
  • System-level latency may not scale exactly with refiner parameter count, because patch cropping, ROI extraction, and consistency-aware inference add fixed overheads beyond the refiner forward pass.
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

4 major / 4 minor

Summary. The paper introduces PatchRefiner V2 (PRV2), a tile-based high-resolution metric depth estimation framework that replaces the heavy refiner branch of PatchRefiner with lightweight encoders. To compensate for the resulting loss of depth-aligned features, the authors propose a Coarse-to-Fine (C2F) module containing Guided Denoising Units (GDUs), a Noisy Pretraining (NP) strategy, and a Scale-and-Shift Invariant Gradient Matching (SSIGM) loss for synthetic-to-real transfer. They evaluate on UnrealStereo4K and on Cityscapes, ScanNet++, and KITTI, and claim that PRV2C achieves a new state of the art on UnrealStereo4K with RMSE 0.884 while being 2.3x faster than PatchRefiner, with improved boundary delineation on real-domain datasets.

Significance. If the claims are validated, the paper makes a useful practical contribution: it shows that a lightweight refiner with a learned denoising module can recover much of the quality of a heavy refiner while drastically reducing parameter count and inference time. The ablations in Tables 4, 6, and 7 consistently show monotonic improvements from end-to-end training, C2F, and NP across three model sizes, which is a genuine strength. The evaluation uses external datasets and external base models, and the use of the previous PatchRefiner model as the pseudo-label teacher is standard self-training rather than circular reasoning. However, two load-bearing issues currently prevent the headline claims from being accepted: the SSIGM loss as written in Eq. (6) is identically zero, and the state-of-the-art comparison in Table 1 relies on reimplemented, 'aligned' baselines whose details are not fully specified and whose margin over the original published PatchRefiner is very small.

major comments (4)
  1. [§3.3, Eq. (6)] The SSIGM loss is defined with R_i = \hat d_i - \hat d^*_i, but Eq. (4) states \hat d^*_i = \hat d_i. Therefore R_i is identically zero and L_ssigm in Eq. (6) is identically zero as written. This makes the SSIGM contribution vacuous on paper and means the real-domain boundary gains reported in Tables 3 and 5 cannot currently be attributed to the proposed loss. Please correct the definition (presumably R_i should be d^*_i - \hat d^*_i) and confirm that the experiments used the corrected formulation.
  2. [§4.3, Table 1] The headline claim that PRV2C sets a new state of the art rests on the 'aligned' baselines PR† and PF†, which are obtained by removing a non-public MiDaS pretraining stage from the refiner or fine branch. The original PatchRefiner row reports RMSE 0.892, while PRV2C reports 0.884, a difference of only 0.008 (about 0.9%); the gap to the aligned PR† is 0.057. Because the alignment procedure is not described in enough detail, no trained models or checkpoints are released, and no multiple-seed variance is reported, the claimed margin over the actual published state of the art is not verifiable and may be within run-to-run variation. Please report results against the original published PatchRefiner and PatchFusion models, provide error bars or multiple runs, and release code/checkpoints.
  3. [§4.3, Table 1 caption] The caption states that #param and T denote the additional parameters and the inference time of the fine or refiner branch only. The coarse branch is shared across compared methods, so the end-to-end inference speedup is smaller than the reported 2.3x. Since the abstract claims that PRV2 outperforms previous methods in both accuracy and speed, the paper should report total end-to-end inference time (coarse plus refiner) or clearly qualify the speed comparison as refiner-branch-only.
  4. [§3.2.3] The Noisy Pretraining strategy replaces input coarse features with random normal noise, and the ablation Table 4 attributes a substantial RMSE improvement to this component. However, the paper does not specify the noise scale, whether the noise is resampled at every optimization step, or how the random features are matched to the resolution of each C2F layer. Without these details, the NP component cannot be reproduced from the manuscript.
minor comments (4)
  1. [Throughout] There are several typos and naming inconsistencies: 'CityScape' should be 'Cityscapes'; 'Midas' should be 'MiDaS'; 'Convnext' should be 'ConvNeXt'; and in §3.3 'predicted depthdi' lacks a space between 'depth' and 'd_i'.
  2. [Table 3] The column header 'Data pl gen' is unclear; please spell out that 'pl' denotes pseudo-labels and 'gen' denotes their online or offline generation.
  3. [Table 5 caption] The notation 'GMSSI' is defined only indirectly; please state explicitly that GMSSI means gradient maps are computed first and then aligned with LSE, in contrast to SSIGM, so that the order-of-operations ablation is self-contained.
  4. [§4.1] The KITTI setup cites Garg et al. for cropping to 352x1216, but the usual Eigen split evaluation uses a specific crop; please clarify whether the reported protocol matches the Eigen split exactly or deviates from it.

Circularity Check

1 steps flagged · score 6.0 of 10

SSIGM loss is identically zero by the paper's own Eqs. (4) and (6), so the claimed boundary-F1 gains reduce by construction; the rest of the framework evaluation is self-contained.

  1. self definitional [Section 3.3, Eqs. (4) and (6)]
    "d* = sd + t, d^hat* = d^hat (4) ... L_ssigm = 1/M sum_i (|grad_x R_i| + |grad_y R_i|), (6) where R_i = d^hat_i - d^hat*_i."

    Eq. (4) defines the aligned pseudo label as unchanged: d^hat*_i = d^hat_i. Eq. (6) then defines the SSIGM residual as R_i = d^hat_i - d^hat*_i, which is identically zero for every pixel. Hence the gradient-matching loss is zero by construction, and the paper's later attribution of boundary improvements ('+17.2% boundary F1 on CityScape w.r.t [36]', Table 3) to this loss is a definitional reduction: the stated objective cannot supply the gradient signal claimed. An unspecified residual involving the aligned prediction, such as d*_i - d^hat*_i, would be needed; as written, the claimed synthetic-to-real transfer gain is not derived from the stated loss.

full rationale

The central speed-accuracy comparison is not circular: PRV2 is evaluated on external UnrealStereo4K ground truth, with external base models (ZoeDepth, Depth Anything V2), and the ablations for C2F and Noisy Pretraining are internal and self-contained. The teacher-student pseudo-labeling uses the authors' own PatchRefiner as teacher, but that is standard self-training rather than a definitional loop, and the real-domain boundary metrics use independent CityScapes labels. The 'aligned version' comparison is a protocol choice that changes the baseline strength, but it is not a derivation that reduces to the paper's own inputs. The one concrete definitional collapse is the SSIGM loss: by Eqs. (4) and (6), every residual R_i is identically zero, so the stated loss is exactly zero and the reported boundary gains attributed to SSIGM are vacuous as written. This affects one of the three headline contributions, so the paper is partially circular rather than entirely self-contained.

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

The paper introduces no new physical entities. Its claims rest on architectural hyperparameters, an empirical pretraining trick, and a loss modification; the main unverified premise is the reliability of the fixed coarse branch and the fairness of the retrained baselines.

free parameters (5)
  • Number of C2F layers N = not stated (ablation uses fixed N)
    Architecture depth chosen by hand; affects feature denoising capacity.
  • Patch size = 540x960 on UnrealStereo4K
    Set for compatibility with PatchFusion and PatchRefiner; other datasets use different sizes.
  • Training epochs = 96 (Noisy Pretraining), 24 (coarse), 48 (end-to-end)
    Chosen without a stated search; affects convergence.
  • Pseudo-label supervision weight lambda = not specified
    Controls the strength of SSIGM loss in the DSD objective.
  • Lightweight encoder choice = MobileNet-Small, EfficientNet-B5, ConvNext-Large
    Models trade off size versus accuracy; the paper evaluates three, none is derived.
assumptions (4)
  • domain assumption The coarse branch provides a reliable base depth estimate Dc.
    Invoked in Section 3.2.1 to justify replacing the heavy refiner; if false, the refinement cannot fix global errors.
  • ad hoc to paper Random normal noise as coarse features during pretraining forces the refiner to learn depth-relevant features.
    Noisy Pretraining (Section 3.2.3) is an empirical heuristic without theoretical justification; the paper only tests it indirectly via ablations.
  • domain assumption Gradient matching on scale-shift aligned depth improves high-frequency boundary learning.
    Borrowed from MegaDepth [32] and validated here by CityScapes boundary F1, but no proof that it generalizes.
  • ad hoc to paper The aligned baseline (removing Midas pretraining) is a fair comparison.
    Table 1 footnote states the modification; if the reimplementation is imperfect, the speed-accuracy claims shift.

how reviews work

0 comments
Cite this review

Pith. "Pith review of PatchRefiner V2: Fast and Lightweight Real-Domain High-Resolution Metric Depth Estimation." pith.science (2026). https://pith.science/paper/WTWOEV7S

@misc{pith2026250101121,
  author       = {Pith},
  title        = {Pith review of: PatchRefiner V2: Fast and Lightweight Real-Domain High-Resolution Metric Depth Estimation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/WTWOEV7S}},
  note         = {Machine review of arXiv:2501.01121}
}
read the original abstract

While current high-resolution depth estimation methods achieve strong results, they often suffer from computational inefficiencies due to reliance on heavyweight models and multiple inference steps, increasing inference time. To address this, we introduce PatchRefiner V2 (PRV2), which replaces heavy refiner models with lightweight encoders. This reduces model size and inference time but introduces noisy features. To overcome this, we propose a Coarse-to-Fine (C2F) module with a Guided Denoising Unit for refining and denoising the refiner features and a Noisy Pretraining strategy to pretrain the refiner branch to fully exploit the potential of the lightweight refiner branch. Additionally, we introduce a Scale-and-Shift Invariant Gradient Matching (SSIGM) loss to enhance synthetic-to-real domain transfer. PRV2 outperforms state-of-the-art depth estimation methods on UnrealStereo4K in both accuracy and speed, using fewer parameters and faster inference. It also shows improved depth boundary delineation on real-world datasets like CityScape, ScanNet++, and KITTI, demonstrating its versatility across domains.

Figures

Figures reproduced from arXiv: 2501.01121 by the authors.

Figure 1
Figure 1. UnrealStereo4K results. PatchRefiner V2 (PRV2) sig￾nificantly outperforms previous high-resolution frameworks. In particular, PRV2C achieves new SOTA RMSE but being 2.3x faster than PR. PRV2M is 9.2x smaller and 10.7x faster than PF. PF and PR are short for PatchFusion [34] and PatchRefiner [36], respectively. We present the comparison of PR and PRV2 in [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. A comparison of (a) PatchRefiner and (b) our proposed [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗
Figure 3
Figure 3. Visualization of F2C input feature maps. We showcase the first 16 channels of the F2C input features. (c) Without the C2F module (setting ③ in Tab. 4), the refiner features are ‘noisy’ and hard to interpret. (d) The C2F module helps denoise the refiner features, leading to clear boundaries and better results. depth estimation [34]. To relax the constraints, initial ef￾forts utilize Guided Depth Super-Resolution (GDS… view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Left: Coarse-to-Fine (C2F) module overview. It processes refiner features in a bottom-to-top manner with N successive C2F layers. Each layer is guided by coarse features with corresponding resolution and outputs denoised features for the Fine-to-Coarse (F2C) module. Ce…
Figure 5
Figure 5. Figure 5: Qualitative Comparison on UnrealStereo4K. We show the depth prediction and corresponding error map, respectively. The qualitative comparisons showcased here indicate our PRV2C outperforms counterparts [4, 36] with sharper edges and lower error around boundaries while a…
Figure 6
Figure 6. Figure 6: Qualitative Comparison on KITTI and ScanNet++. Our PVR2 can consistently boost high-resolution depth estimation with various base models (ZoeDepth, DAV2) on various real-domain datasets. Zoom in to better perceive details near boundaries. Method Data pl gen Scale Bound…

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

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

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

    cs.CV 2026-07 unverdicted novelty 6.0 of 10

    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

78 extracted references · 61 canonical work pages · cited by 1 Pith paper

  1. [1]

    Beit: Bert pre-training of image transformers

    Hangbo Bao, Li Dong, Songhao Piao, and Furu Wei. Beit: Bert pre-training of image transformers. arXiv preprint arXiv:2106.08254, 2021. 2

    arXiv 2021

  2. [2]

    Adabins: Depth estimation using adaptive bins

    Shariq Farooq Bhat, Ibraheem Alhashim, and Peter Wonka. Adabins: Depth estimation using adaptive bins. In CVPR, pages 4009–4018, 2021. 2, 7

    work page 2021

  3. [3]

    Localbins: Improving depth estimation by learning local dis- tributions

    Shariq Farooq Bhat, Ibraheem Alhashim, and Peter Wonka. Localbins: Improving depth estimation by learning local dis- tributions. In European Conference on Computer Vision , pages 480–496. Springer, 2022. 2

    work page 2022

  4. [4]

    Zoedepth: Zero-shot trans- fer by combining relative and metric depth

    Shariq Farooq Bhat, Reiner Birkl, Diana Wofk, Peter Wonka, and Matthias M ¨uller. Zoedepth: Zero-shot trans- fer by combining relative and metric depth. arXiv preprint arXiv:2302.12288, 2023. 1, 2, 4, 5, 6, 7

    arXiv 2023

  5. [5]

    In- structpix2pix: Learning to follow image editing instructions

    Tim Brooks, Aleksander Holynski, and Alexei A Efros. In- structpix2pix: Learning to follow image editing instructions. In CVPR, pages 18392–18402, 2023. 5

    work page 2023

  6. [6]

    On the over-smoothing problem of cnn based disparity estimation

    Chuangrong Chen, Xiaozhi Chen, and Hui Cheng. On the over-smoothing problem of cnn based disparity estimation. In ICCV, pages 8997–9005, 2019. 7

    work page 2019

  7. [7]

    Progressive feature alignment for unsupervised do- main adaptation

    Chaoqi Chen, Weiping Xie, Wenbing Huang, Yu Rong, Xinghao Ding, Yue Huang, Tingyang Xu, and Junzhou Huang. Progressive feature alignment for unsupervised do- main adaptation. In CVPR, pages 627–636, 2019. 3

    work page 2019

  8. [8]

    Learning con- tinuous image representation with local implicit image func- tion

    Yinbo Chen, Sifei Liu, and Xiaolong Wang. Learning con- tinuous image representation with local implicit image func- tion. In CVPR, pages 8628–8638, 2021. 3

    work page 2021

Show all 78 references
  1. [9]

    Crdoco: Pixel-level domain transfer with cross- domain consistency

    Yun-Chun Chen, Yen-Yu Lin, Ming-Hsuan Yang, and Jia- Bin Huang. Crdoco: Pixel-level domain transfer with cross- domain consistency. In CVPR, pages 1791–1800, 2019. 3

    2019

  2. [10]

    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. In CVPR, pages 3213–3223, 2016. 2, 6, 9

    2016

  3. [11]

    Scannet: Richly-annotated 3d reconstructions of indoor scenes

    Angela Dai, Angel X Chang, Manolis Savva, Maciej Hal- ber, Thomas Funkhouser, and Matthias Nießner. Scannet: Richly-annotated 3d reconstructions of indoor scenes. In CVPR, pages 5828–5839, 2017. 2

    2017

  4. [12]

    Learn- ing graph regularisation for guided super-resolution

    Riccardo De Lutio, Alexander Becker, Stefano D’Aronco, Stefania Russo, Jan D Wegner, and Konrad Schindler. Learn- ing graph regularisation for guided super-resolution. In CVPR, pages 1979–1988, 2022. 6

    1979

  5. [13]

    Imagenet: A large-scale hierarchical image database

    Jia Deng, Wei Dong, Richard Socher, Li-Jia Li, Kai Li, and Li Fei-Fei. Imagenet: A large-scale hierarchical image database. In CVPR, pages 248–255. Ieee, 2009. 2, 8

    2009

  6. [14]

    Soft labels for ordinal regres- sion

    Raul Diaz and Amit Marathe. Soft labels for ordinal regres- sion. In CVPR, pages 4738–4747, 2019. 2

    2019

  7. [15]

    An image is worth 16x16 words: Trans- formers for image recognition at scale

    Alexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn, Xiaohua Zhai, Thomas Unterthiner, Mostafa Dehghani, Matthias Minderer, Georg Heigold, Syl- vain Gelly, et al. An image is worth 16x16 words: Trans- formers for image recognition at scale. arXiv preprint a...

    2010 arXiv

  8. [16]

    Depth map prediction from a single image using a multi-scale deep net- work

    David Eigen, Christian Puhrsch, and Rob Fergus. Depth map prediction from a single image using a multi-scale deep net- work. NeurIPS, 27, 2014. 1, 2, 7, 9

    2014

  9. [17]

    Con- trastive learning for depth prediction

    Rizhao Fan, Matteo Poggi, and Stefano Mattoccia. Con- trastive learning for depth prediction. In CVPRW, pages 3225–3236, 2023. 2

    2023

  10. [18]

    Deep ordinal regression net- work for monocular depth estimation

    Huan Fu, Mingming Gong, Chaohui Wang, Kayhan Bat- manghelich, and Dacheng Tao. Deep ordinal regression net- work for monocular depth estimation. InCVPR, pages 2002– 2011, 2018. 2

    2002

  11. [19]

    Unsupervised cnn for single view depth estimation: Geom- etry to the rescue

    Ravi Garg, Vijay Kumar Bg, Gustavo Carneiro, and Ian Reid. Unsupervised cnn for single view depth estimation: Geom- etry to the rescue. In Computer Vision–ECCV 2016: 14th European Conference, Amsterdam, The Netherlands, Octo- ber 11-14, 2016, Proceedings, Part VIII 14, pages 7...

    2016

  12. [20]

    Are we ready for autonomous driving? the kitti vision benchmark suite

    Andreas Geiger, Philip Lenz, and Raquel Urtasun. Are we ready for autonomous driving? the kitti vision benchmark suite. In CVPR, pages 3354–3361. IEEE, 2012. 9

    2012

  13. [21]

    Vision meets robotics: The kitti dataset

    Andreas Geiger, Philip Lenz, Christoph Stiller, and Raquel Urtasun. Vision meets robotics: The kitti dataset. IJRR, 32 (11):1231–1237, 2013. 2

    2013

  14. [22]

    Digging into self-supervised monocular depth estimation

    Cl ´ement Godard, Oisin Mac Aodha, Michael Firman, and Gabriel J Brostow. Digging into self-supervised monocular depth estimation. In ICCV, pages 3828–3838, 2019. 2

    2019

  15. [23]

    Mask r-cnn

    Kaiming He, Georgia Gkioxari, Piotr Doll ´ar, and Ross Gir- shick. Mask r-cnn. In ICCV, pages 2961–2969, 2017. 3

    2017

  16. [24]

    Searching for mo- bilenetv3

    Andrew Howard, Mark Sandler, Grace Chu, Liang-Chieh Chen, Bo Chen, Mingxing Tan, Weijun Wang, Yukun Zhu, Ruoming Pang, Vijay Vasudevan, et al. Searching for mo- bilenetv3. In CVPR, pages 1314–1324, 2019. 2

    2019

  17. [25]

    Depth map super-resolution by deep multi-scale guidance

    Tak-Wai Hui, Chen Change Loy, and Xiaoou Tang. Depth map super-resolution by deep multi-scale guidance. In ECCV, pages 353–369. Springer, 2016. 3

    2016

  18. [26]

    Repurpos- ing diffusion-based image generators for monocular depth estimation

    Bingxin Ke, Anton Obukhov, Shengyu Huang, Nando Met- zger, Rodrigo Caye Daudt, and Konrad Schindler. Repurpos- ing diffusion-based image generators for monocular depth estimation. In CVPR, pages 9492–9502, 2024. 2

    2024

  19. [27]

    Segment any- thing

    Alexander Kirillov, Eric Mintun, Nikhila Ravi, Hanzi Mao, Chloe Rolland, Laura Gustafson, Tete Xiao, Spencer White- head, Alexander C Berg, Wan-Yen Lo, et al. Segment any- thing. In ICCV, pages 4015–4026, 2023. 3 10

    2023

  20. [28]

    Sharingan: Combining synthetic and real data for unsupervised geome- try estimation

    PNVR Koutilya, Hao Zhou, and David Jacobs. Sharingan: Combining synthetic and real data for unsupervised geome- try estimation. In CVPR, page 5, 2020. 3

    2020

  21. [29]

    Adadepth: Unsupervised content congruent adaptation for depth estimation

    Jogendra Nath Kundu, Phani Krishna Uppala, Anuj Pahuja, and R Venkatesh Babu. Adadepth: Unsupervised content congruent adaptation for depth estimation. In CVPR, pages 2656–2665, 2018. 3

    2018

  22. [30]

    Multi-loss rebalancing al- gorithm for monocular depth estimation

    Jae-Han Lee and Chang-Su Kim. Multi-loss rebalancing al- gorithm for monocular depth estimation. In ECCV, pages 785–801. Springer, 2020. 2

    2020

  23. [31]

    From big to small: Multi-scale local planar guidance for monocular depth estimation

    Jin Han Lee, Myung-Kyu Han, Dong Wook Ko, and Il Hong Suh. From big to small: Multi-scale local planar guidance for monocular depth estimation. arXiv preprint arXiv:1907.10326, 2019. 7

    1907 arXiv

  24. [32]

    Megadepth: Learning single- view depth prediction from internet photos

    Zhengqi Li and Noah Snavely. Megadepth: Learning single- view depth prediction from internet photos. In CVPR, pages 2041–2050, 2018. 2, 3, 6, 9

    2018

  25. [33]

    Binsformer: Revisiting adaptive bins for monocular depth estimation

    Zhenyu Li, Xuyang Wang, Xianming Liu, and Junjun Jiang. Binsformer: Revisiting adaptive bins for monocular depth estimation. arXiv preprint arXiv:2204.00987, 2022. 1, 2

    2022 arXiv

  26. [34]

    Patch- fusion: An end-to-end tile-based framework for high- resolution monocular metric depth estimation.arXiv preprint arXiv:2312.02284, 2023

    Zhenyu Li, Shariq Farooq Bhat, and Peter Wonka. Patch- fusion: An end-to-end tile-based framework for high- resolution monocular metric depth estimation.arXiv preprint arXiv:2312.02284, 2023. 1, 3, 4, 5, 6, 7, 9

    2023 arXiv

  27. [35]

    Depthformer: Exploiting long-range correlation and local in- formation for accurate monocular depth estimation.Machine Intelligence Research, pages 1–18, 2023

    Zhenyu Li, Zehui Chen, Xianming Liu, and Junjun Jiang. Depthformer: Exploiting long-range correlation and local in- formation for accurate monocular depth estimation.Machine Intelligence Research, pages 1–18, 2023. 2

    2023

  28. [36]

    Patchre- finer: Leveraging synthetic data for real-domain high- resolution monocular metric depth estimation.arXiv preprint arXiv:2406.06679, 2024

    Zhenyu Li, Shariq Farooq Bhat, and Peter Wonka. Patchre- finer: Leveraging synthetic data for real-domain high- resolution monocular metric depth estimation.arXiv preprint arXiv:2406.06679, 2024. 1, 2, 3, 4, 5, 6, 7, 8, 9

    2024 arXiv

  29. [37]

    Refinenet: Multi-path refinement networks for high- resolution semantic segmentation

    Guosheng Lin, Anton Milan, Chunhua Shen, and Ian Reid. Refinenet: Multi-path refinement networks for high- resolution semantic segmentation. In CVPR, pages 1925– 1934, 2017. 2, 4

    1925

  30. [38]

    Single image depth prediction made better: A multivariate gaussian take

    Ce Liu, Suryansh Kumar, Shuhang Gu, Radu Timofte, and Luc Van Gool. Single image depth prediction made better: A multivariate gaussian take. In CVPR, pages 17346–17356,

  31. [39]

    Zero-1-to-3: Zero-shot one image to 3d object

    Ruoshi Liu, Rundi Wu, Basile Van Hoorick, Pavel Tok- makov, Sergey Zakharov, and Carl V ondrick. Zero-1-to-3: Zero-shot one image to 3d object. In ICCV, pages 9298– 9309, 2023. 5

    2023

  32. [40]

    Desc: Domain adaptation for depth estimation via semantic con- sistency

    Adrian Lopez-Rodriguez and Krystian Mikolajczyk. Desc: Domain adaptation for depth estimation via semantic con- sistency. International Journal of Computer Vision, 131(3): 752–771, 2023. 3

    2023

  33. [41]

    Guided depth super-resolution by deep anisotropic diffusion

    Nando Metzger, Rodrigo Caye Daudt, and Konrad Schindler. Guided depth super-resolution by deep anisotropic diffusion. In CVPR, pages 18237–18246, 2023. 3

    2023

  34. [42]

    Boosting monocular depth estima- tion models to high-resolution via content-adaptive multi- resolution merging

    S Mahdi H Miangoleh, Sebastian Dille, Long Mai, Sylvain Paris, and Yagiz Aksoy. Boosting monocular depth estima- tion models to high-resolution via content-adaptive multi- resolution merging. In CVPR, pages 9685–9694, 2021. 3, 6

    2021

  35. [43]

    Nerf: Representing scenes as neural radiance fields for view syn- thesis

    Ben Mildenhall, Pratul P Srinivasan, Matthew Tancik, Jonathan T Barron, Ravi Ramamoorthi, and Ren Ng. Nerf: Representing scenes as neural radiance fields for view syn- thesis. Communications of the ACM , 65(1):99–106, 2021. 3

    2021

  36. [44]

    Dinov2: Learning robust visual features without supervision

    Maxime Oquab, Timoth ´ee Darcet, Th ´eo 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. 2

    2023 arXiv

  37. [45]

    pix2gestalt: Amodal segmentation by synthesizing wholes

    Ege Ozguroglu, Ruoshi Liu, D ´ıdac Sur´ıs, Dian Chen, Achal Dave, Pavel Tokmakov, and Carl V ondrick. pix2gestalt: Amodal segmentation by synthesizing wholes. In CVPR, pages 3931–3940. IEEE Computer Society, 2024. 5

    2024

  38. [46]

    A closer look at self-training for zero-label semantic segmentation

    Giuseppe Pastore, Fabio Cermelli, Yongqin Xian, Massimil- iano Mancini, Zeynep Akata, and Barbara Caputo. A closer look at self-training for zero-label semantic segmentation. In CVPR, pages 2693–2702, 2021. 3

    2021

  39. [47]

    Exploiting pseudo labels in a self-supervised learning framework for improved monocular depth estimation

    Andra Petrovai and Sergiu Nedevschi. Exploiting pseudo labels in a self-supervised learning framework for improved monocular depth estimation. In CVPR, pages 1578–1588,

  40. [48]

    idisc: Internal discretization for monocular depth estimation

    Luigi Piccinelli, Christos Sakaridis, and Fisher Yu. idisc: Internal discretization for monocular depth estimation. In CVPR, pages 21477–21487, 2023. 6, 7

    2023

  41. [49]

    Boost- ing instance segmentation with synthetic data: A study to overcome the limits of real world data sets

    Florentin Poucin, Andrea Kraus, and Martin Simon. Boost- ing instance segmentation with synthetic data: A study to overcome the limits of real world data sets. In Int. Conf. Comput. Vis. Worksh., pages 945–953, 2021. 3

    2021

  42. [50]

    The simple and efficient semi-supervised learning method for deep neural networks

    Dong-Hyun Lee Pseudo-Label. The simple and efficient semi-supervised learning method for deep neural networks. In ICML 2013 Workshop: Challenges in Representation Learning, pages 1–6, 2013. 3

    2013

  43. [51]

    Mobilenetv4- universal models for the mobile ecosystem

    Danfeng Qin, Chas Leichner, Manolis Delakis, Marco Fornoni, Shixin Luo, Fan Yang, Weijun Wang, Colby Ban- bury, Chengxi Ye, Berkin Akin, et al. Mobilenetv4- universal models for the mobile ecosystem. arXiv preprint arXiv:2404.10518, 2024. 2, 7

    2024 arXiv

  44. [52]

    High-resolution syn- thetic rgb-d datasets for monocular depth estimation

    Aakash Rajpal, Noshaba Cheema, Klaus Illgner-Fehns, Philipp Slusallek, and Sunil Jaiswal. High-resolution syn- thetic rgb-d datasets for monocular depth estimation. In CVPR, pages 1188–1198, 2023. 3

    2023

  45. [53]

    Vi- sion transformers for dense prediction

    Ren ´e Ranftl, Alexey Bochkovskiy, and Vladlen Koltun. Vi- sion transformers for dense prediction. In ICCV, pages 12179–12188, 2021. 2

    2021

  46. [54]

    Towards robust monocular depth estimation: Mixing datasets for zero-shot cross-dataset transfer

    Ren ´e Ranftl, Katrin Lasinger, David Hafner, Konrad Schindler, and Vladlen Koltun. Towards robust monocular depth estimation: Mixing datasets for zero-shot cross-dataset transfer. IEEE TPAMI, 44(3), 2022. 2, 3, 4, 6, 7, 9

    2022

  47. [55]

    Hypersim: A photorealistic syn- thetic dataset for holistic indoor scene understanding

    Mike Roberts, Jason Ramapuram, Anurag Ranjan, Atulit Kumar, Miguel Angel Bautista, Nathan Paczan, Russ Webb, and Joshua M Susskind. Hypersim: A photorealistic syn- thetic dataset for holistic indoor scene understanding. In ICCV, pages 10912–10922, 2021. 2

    2021

  48. [56]

    High-resolution image syn- thesis with latent diffusion models

    Robin Rombach, Andreas Blattmann, Dominik Lorenz, Patrick Esser, and Bj¨orn Ommer. High-resolution image syn- thesis with latent diffusion models. In CVPR, pages 10684– 10695, 2022. 2 11

    2022

  49. [57]

    U-net: Convolutional networks for biomedical image segmentation,

    Olaf Ronneberger, Philipp Fischer, and Thomas Brox. U-net: Convolutional networks for biomedical image segmentation,

  50. [58]

    Asymmetric tri-training for unsupervised domain adaptation

    Kuniaki Saito, Yoshitaka Ushiku, and Tatsuya Harada. Asymmetric tri-training for unsupervised domain adaptation. In International Conference on Machine Learning , pages 2988–2997. PMLR, 2017. 3

    2017

  51. [59]

    High-resolution stereo datasets with subpixel-accurate ground truth

    Daniel Scharstein, Heiko Hirschm ¨uller, York Kitajima, Greg Krathwohl, Nera Ne ˇsi´c, Xi Wang, and Porter West- ling. High-resolution stereo datasets with subpixel-accurate ground truth. In Pattern Recognition: 36th German Confer- ence, GCPR 2014, M¨unster, Germany, September...

    2014

  52. [60]

    A multi-view stereo benchmark with high- resolution images and multi-camera videos

    Thomas Schops, Johannes L Schonberger, Silvano Galliani, Torsten Sattler, Konrad Schindler, Marc Pollefeys, and An- dreas Geiger. A multi-view stereo benchmark with high- resolution images and multi-camera videos. In CVPR, pages 3260–3269, 2017. 2, 9

    2017

  53. [61]

    Mm-tta: multi-modal test-time adaptation for 3d se- mantic segmentation

    Inkyu Shin, Yi-Hsuan Tsai, Bingbing Zhuang, Samuel Schulter, Buyu Liu, Sparsh Garg, In So Kweon, and Kuk-Jin Yoon. Mm-tta: multi-modal test-time adaptation for 3d se- mantic segmentation. In CVPR, pages 16928–16937, 2022. 3

    2022

  54. [62]

    Indoor segmentation and support inference from rgbd images

    Nathan Silberman, Derek Hoiem, Pushmeet Kohli, and Rob Fergus. Indoor segmentation and support inference from rgbd images. In ECCV, pages 746–760. Springer, 2012. 2, 7, 9

    2012

  55. [63]

    Sun rgb-d: A rgb-d scene understanding benchmark suite

    Shuran Song, Samuel P Lichtenberg, and Jianxiong Xiao. Sun rgb-d: A rgb-d scene understanding benchmark suite. In CVPR, pages 567–576, 2015. 9

    2015

  56. [64]

    Mingxing Tan and Quoc V . Le. Efficientnet: Rethinking model scaling for convolutional neural networks. CoRR, abs/1905.11946, 2019. 2, 7

    1905 arXiv

  57. [65]

    Smd-nets: Stereo mixture density networks

    Fabio Tosi, Yiyi Liao, Carolin Schmitt, and Andreas Geiger. Smd-nets: Stereo mixture density networks. In CVPR, pages 8942–8952, 2021. 2, 6, 7, 9

    2021

  58. [66]

    A survey on semi-supervised learning

    Jesper E Van Engelen and Holger H Hoos. A survey on semi-supervised learning. Machine learning , 109(2):373– 440, 2020. 3

    2020

  59. [67]

    Image quality assessment: from error visibility to structural similarity

    Zhou Wang, Alan C Bovik, Hamid R Sheikh, and Eero P Simoncelli. Image quality assessment: from error visibility to structural similarity. IEEE TIP, 13(4):600–612, 2004. 9

    2004

  60. [68]

    Monocular relative depth per- ception with web stereo data supervision

    Ke Xian, Chunhua Shen, Zhiguo Cao, Hao Lu, Yang Xiao, Ruibo Li, and Zhenbo Luo. Monocular relative depth per- ception with web stereo data supervision. In CVPR, pages 311–320, 2018. 2

    2018

  61. [69]

    Structure-guided ranking loss for single image depth prediction

    Ke Xian, Jianming Zhang, Oliver Wang, Long Mai, Zhe Lin, and Zhiguo Cao. Structure-guided ranking loss for single image depth prediction. In CVPR, pages 611–620, 2020. 2, 3, 7

    2020

  62. [70]

    Transformer-based attention networks for con- tinuous pixel-wise prediction

    Guanglei Yang, Hao Tang, Mingli Ding, Nicu Sebe, and Elisa Ricci. Transformer-based attention networks for con- tinuous pixel-wise prediction. In ICCV, pages 16269–16279,

  63. [71]

    Depth anything: Unleash- ing the power of large-scale unlabeled data

    Lihe Yang, Bingyi Kang, Zilong Huang, Xiaogang Xu, Ji- ashi Feng, and Hengshuang Zhao. Depth anything: Unleash- ing the power of large-scale unlabeled data. arXiv preprint arXiv:2401.10891, 2024. 1, 2, 3, 4, 7, 9

    2024 arXiv

  64. [72]

    Depth any- thing v2

    Lihe Yang, Bingyi Kang, Zilong Huang, Zhen Zhao, Xiao- gang Xu, Jiashi Feng, and Hengshuang Zhao. Depth any- thing v2. arXiv preprint arXiv:2406.09414 , 2024. 1, 2, 4, 6

    2024 arXiv

  65. [73]

    Scannet++: A high-fidelity dataset of 3d indoor scenes

    Chandan Yeshwanth, Yueh-Cheng Liu, Matthias Nießner, and Angela Dai. Scannet++: A high-fidelity dataset of 3d indoor scenes. In ICCV, pages 12–22, 2023. 2, 9

    2023

  66. [74]

    Adding conditional control to text-to-image diffusion models

    Lvmin Zhang, Anyi Rao, and Maneesh Agrawala. Adding conditional control to text-to-image diffusion models. In ICCV, pages 3836–3847, 2023. 1

    2023

  67. [75]

    Geometry-aware symmetric domain adaptation for monocular depth estimation

    Shanshan Zhao, Huan Fu, Mingming Gong, and Dacheng Tao. Geometry-aware symmetric domain adaptation for monocular depth estimation. In CVPR, pages 9788–9798,

  68. [76]

    Discrete cosine transform network for guided depth map super-resolution

    Zixiang Zhao, Jiangshe Zhang, Shuang Xu, Zudi Lin, and Hanspeter Pfister. Discrete cosine transform network for guided depth map super-resolution. In CVPR, pages 5697– 5707, 2022. 3

    2022

  69. [77]

    T2net: Synthetic-to-realistic translation for solving single-image depth estimation tasks

    Chuanxia Zheng, Tat-Jen Cham, and Jianfei Cai. T2net: Synthetic-to-realistic translation for solving single-image depth estimation tasks. In ECCV, pages 767–783, 2018. 3

    2018

  70. [78]

    Guided depth map super-resolution: A survey

    Zhiwei Zhong, Xianming Liu, Junjun Jiang, Debin Zhao, and Xiangyang Ji. Guided depth map super-resolution: A survey. ACM Computing Surveys, 2023. 3 12

    2023

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.