Pith. sign in

REVIEW 4 major objections 6 minor 59 references

FiffDepth: Feed-forward Transformation of Diffusion-Based Generators for Detailed Depth Estimation

T0 review · 4 major / 6 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read FiffDepth claims that a pre-trained Stable Diffusion image generator, read at t=0 and filtered at t=-1, becomes a one-step depth estimator that matches or beats larger zero-shot models on detail and generalization.

desk verdict Solid engineering paper that turns Stable Diffusion into a fast one-step depth estimator; the core recipe is plausible but two unvalidated pieces (t=-1 filter, shared-weight interference) need attention. read the letter →

arxiv 2412.00671 v2 pith:3GGOPQUA submitted 2024-12-01 cs.CV

Yunpeng Bai , Qixing Huang This is my paper · ORCID
classification cs.CV
keywords monoculardepthestimationdiffusionpriorsfeed-forwardStablezero-shotgeneralizationDINOv2distillationboundarieslatent-space
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

FiffDepth claims that a pre-trained Stable Diffusion image generator can be turned into a deterministic, one-step feed-forward depth estimator without losing the fine detail that makes generative models attractive. The key move is to evaluate the diffusion denoiser directly at time step $t=0$ for depth prediction and at $t=-1$ as a learnable filter, both outside its original training range, while keeping the original generative denoising trajectory alive during fine-tuning by blending image and depth latents. Because the $t=-1$ filter is supervised with pseudo-labels from a DINOv2-based model on unlabeled real images, the model learns DINOv2-style robustness on low-frequency depth while retaining generative high-frequency detail. If the claim holds, detailed and generalizable monocular depth can be obtained from roughly 274K training images in one fast inference pass.

What carries the argument

The central object is the pre-trained Stable Diffusion denoiser itself, reused as a feed-forward network at two nonstandard timesteps: $t=0$ gives the depth latent and $t=-1$ gives a learnable filter output matched to DINOv2 pseudo-labels. The trajectory-preserving objective diffuses a blended latent $b_0 = \gamma x_0 + (1-\gamma) d_0$ under the standard noise schedule and trains with v-prediction, so the model keeps generative image features while learning depth. The final objective combines latent-space MiDaS-style MAE and gradient-matching losses at $t=0$ (synthetic ground truth) and at $t=-1$ (DINOv2 pseudo-labels from real images) with the trajectory loss.

What would settle it

Run the released model with the timestep input clamped to the nearest in-range values and compare depth accuracy, detail, and boundary scores; if performance barely changes, the claimed $t=0$/$t=-1$ behavior is not what carries the result. Alternatively, compute the high-frequency energy of $d_0$ and $d_{-1}$ on unseen real images: if $d_{-1}$ is not actually a low-pass-filtered version of $d_0$, the learnable-filter distillation story fails.

Watch

Extended reading notes

Core claim

On its own terms, the paper's discovery is that the denoising network of a latent diffusion model, evaluated at a single out-of-range timestep, is already a strong deterministic depth estimator, and that a second out-of-range evaluation acts as a learnable filter without adding a separate network. Equation (1) defines the depth latent as $d_0 = \hat{\epsilon}_\theta(x_0, t=0)$, a single forward pass on the RGB image's latent, decoded by the Stable Diffusion VAE. Equation (4) defines $d_{-1} = \hat{\epsilon}_\theta(d_0, t=-1)$, a filtered version that resembles DINOv2's low-frequency predictions; supervising only $d_{-1}$ with DINOv2 pseudo-labels transfers generalization to real images without erasing the details in $d_0$. The paper argues that this two-step scheme, together with the blended training target $b_0 = \gamma x_0 + (1-\gamma) d_0$ and a v-prediction trajectory loss, lets the model match or beat prior zero-shot depth estimators on NYUv2, KITTI, ScanNet, ETH3D, DIODE, and DA-2K, and outperform them on boundary accuracy.

Load-bearing premise

The paper's load-bearing premise is that the fine-tuned denoiser remains well-behaved at timesteps $t=0$ and $t=-1$, which lie outside the $t \in \{1,\ldots,1000\}$ range it was trained on; the paper offers no separate evidence for this extrapolation beyond the final benchmark numbers.

Editorial extensions

If this is right

  • Inference becomes a single deterministic forward pass, about 0.092 s per 512×512 image in the paper's timing table, roughly 10× faster than diffusion-sampling depth models.
  • Training uses only 74K synthetic images plus 0.2M unlabeled real images, versus about 62.6M images for the strongest comparison model, so cheap pseudo-labels may replace massive depth annotation.
  • Detail and generalization no longer trade off: the $t=0$ output keeps high-frequency structure while the $t=-1$ branch injects DINOv2-style low-frequency robustness on real images.
  • Boundary sharpness improves: the method reports higher F1 and boundary recall than prior state-of-the-art on Sintel, Spring, iBims, AM-2k, P3M, and DIS-5k.
  • The same diffusion backbone plays both predictor and filter, so the robustness transfer adds no extra network parameters.

Reading between the lines

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

  • A testable general principle suggested by the method: the timestep index of a diffusion denoiser can serve as a controllable detail dial, with out-of-range steps encoding coarser, more robust structure; the same $t=0$/$t=-1$ two-step scheme could transfer to surface normals, segmentation, or other dense tasks.
  • The blended target $\gamma x_0 + (1-\gamma) d_0$ interpolates between image and task latents, and the paper does not explore how varying $\gamma$ changes the trade-off between generative fidelity and task-specific accuracy; a sweep would expose whether the choice $\gamma=0.5$ is critical.
  • If the load-bearing extrapolation is correct, the feed-forward trick should work on other pre-trained latent diffusion models with minimal adaptation, giving a cheap way to probe which generative priors carry geometric structure.
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 / 6 minor

Summary. The paper presents FiffDepth, a monocular depth estimator obtained by fine-tuning a pre-trained Stable Diffusion U-Net so that it runs as a deterministic one-step feed-forward network. At inference, the model computes d0 = epsilon_theta(x0, t=0) from the image latent and decodes it with the SD VAE. Training preserves the original diffusion trajectory on a blended image-depth target (Eqs. (2)-(3)), adds depth losses on d0 from synthetic data, and applies a second evaluation d_-1 = epsilon_theta(d0, t=-1) supervised by Depth Anything v2-Large pseudo-labels on real images (Eqs. (4)-(7)). The paper reports state-of-the-art or near state-of-the-art zero-shot affine-invariant depth accuracy on NYUv2, KITTI, ETH3D, ScanNet, DIODE, and DA-2K, better boundary metrics than Depth Pro and DAv2 in Table 2, and faster inference than multi-step diffusion MDE methods.

Significance. If the results hold, FiffDepth is a useful practical contribution: it converts a text-to-image diffusion model into a single-step dense predictor with competitive accuracy and detail, uses far less training data than Depth Anything v2, and its deterministic inference is much cheaper than Marigold-style multi-step generative depth estimation. The boundary-metric results in Table 2 are independent evidence for the detail claim, and the release of source code is a further strength. However, the core mechanisms, namely extrapolating the denoiser to t=0 and t=-1, the learned-filter interpretation, and the generalization claim on DA-2K, are not yet rigorously validated, so the contribution is currently under-supported relative to the strength of the claims.

major comments (4)
  1. [Sections 3.1 and 3.3, Eqs. (1) and (4)] The central feed-forward design evaluates the denoiser at t=0 for d0 and at t=-1 for the filter, although the diffusion training objective Eq. (3) is defined only for t in {1,...,T}. The paper gives no evidence that the timestep embedding extrapolates sensibly to these values, and it does not compare against in-range alternatives such as t=1 for d0 or a conventional low-pass filter applied to d0 in place of Eq. (4). The claim that d_-1 is a learned low-pass filter rests on the qualitative Figure 3 and on benchmark numbers, which could be explained by other mechanisms. Please add a quantitative comparison that includes an in-range timestep and a non-learned filter baseline, and provide spectral or correspondence analysis of d_-1 if possible.
  2. [Section 4.2, Table 1 (DA-2K column); Section 4.1] The model is trained to regress Depth Anything v2-Large pseudo-labels at d_-1 (Eq. (4)), and DAv2-L itself scores 97.1 on DA-2K. Reporting the same 97.1 DA-2K accuracy is therefore not evidence of independent zero-shot generalization; it is partly a check of how faithfully the student imitates the teacher. Please either remove DA-2K from the generalization comparison, report it explicitly as a distillation-fidelity measure, or provide zero-shot evidence on benchmarks whose labels were not produced by DAv2-L and are not part of its training or evaluation distribution.
  3. [Section 3.4, Eq. (7) with Eq. (4)] Because d_-1 is obtained by applying the same network epsilon_theta to d0, gradients from the d_-1 losses can flow into d0 through the shared parameters unless a stop-gradient is inserted. The paper does not state whether d0 is detached, so the claim in Section 3.3 that the DAv2-L supervision transfers robustness without interfering with the detailed features in d0 is not supported. Please specify the gradient routing and include an ablation comparing training with and without detaching d0 in Eq. (4).
  4. [Section 4.2, Ablation studies] The ablation evidence in Fig. 7 is qualitative only, even though the paper's main contributions are the trajectory-preserving training and the d_-1 filter distillation. Without quantitative metrics such as AbsRel, delta1, or boundary F1 for each removed component, the reader cannot determine which components drive the reported gains or whether Eq. (4) is necessary at all. Please add a quantitative ablation table covering w/o L_k, w/o blend, w/o DINO supervision, DINO supervision on d0, and replacement of Eq. (4) with a non-learned low-pass filter.
minor comments (6)
  1. [Section 4.1] Please specify the exact Stable Diffusion checkpoint and version, the VAE, whether the whole U-Net or a subset is fine-tuned, the optimizer, learning rate, batch size, resolution, number of iterations, and how the LAION-Art subset is sampled; this is needed to reproduce the 274K training-data claim.
  2. [Equations (5)-(7)] The roles of d and d* are swapped between the definition of the loss, where d is ground truth and d* is prediction, and the sentence after Eq. (7), where d*_0 is ground truth and d*_-1 is pseudo-label; please make the notation consistent.
  3. [Sections 3.3 and 4.1] The teacher model is Depth Anything v2-Large, not DINOv2 itself; please use consistent terminology and clarify which component, DINOv2 features or DAv2 depth pseudo-labels, is actually being distilled.
  4. [Table 2 and boundary-metric paragraph] The computation of F1 and boundary recall is only referenced to Depth Pro; please state the exact preprocessing, alignment, and threshold protocol in the paper or supplement.
  5. [Table 3] The running-time numbers should state the input resolution, GPU, and software versions used for each method, and whether all methods run at the same resolution with their default settings; otherwise the comparison is hard to interpret.
  6. [Throughout] Please fix typographical issues such as 'fintuned' in Section 3.3, 'performe' in the Table 3 caption, and clarify the 'blend' symbol in Figure 2, which is not defined in the caption.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the core depth predictor is supervised by external synthetic ground truth and evaluated on independent benchmarks; the DAv2 distillation is an explicit training objective rather than a hidden re-labeling of the target claim.

full rationale

The derivation chain is self-contained with respect to its major claims. The t=0 prediction d0 = eps_theta(x0,0) (Eq. 1) is trained with L_MAE and L_GM against ground-truth depth from Hypersim and Virtual KITTI (Eqs. 5-7), datasets that are disjoint from the NYUv2, KITTI, ScanNet, ETH3D, and DIODE evaluation sets; those benchmark results are therefore not forced by construction. The t=-1 filter d_{-1} = eps_theta(d0,-1) (Eq. 4) is supervised by Depth Anything v2-L pseudo-labels on LAION-Art real images, and the paper explicitly states this distillation is how DINOv2 generalization is transferred. Reporting DA-2K accuracy equal to the teacher is a plausible consequence of distillation, but the paper does not define DA-2K accuracy as agreement with those pseudo-labels, and the boundary/detail metrics in Table 2 are measured against independent ground-truth depth, matting, and segmentation labels. The use of t=0 and t=-1 outside the DDPM training range is an extrapolation that is not theoretically justified, and the paper does not state whether d0 is detached when computing d_{-1}; these are unverified design and correctness risks, not equation-level circular reductions. There is also a practical limitation noted by the authors (DAv2-Giant weights unavailable), but no load-bearing self-citation chain or renamed-known-result step appears in the manuscript.

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

The central claim rests mainly on hand-chosen hyperparameters and two strong domain assumptions: that negative timesteps can be repurposed, and that a DINOv2-based teacher provides reliable low-frequency supervision. No new physical entities are introduced.

free parameters (4)
  • gamma (blend factor) = 0.5
    Controls tradeoff between image latent and depth latent in Eq. (2); set without a reported search, and central to trajectory preservation.
  • lambda_MAE, lambda_GM, lambda_k (loss weights) = 1, 0.5, 0.2
    Weights in Eq. (7); hand-picked with no sensitivity analysis, and they determine the balance between detail and teacher distillation.
  • t = -1 filter timestep = -1
    The filter in Eq. (4) is implemented as a network evaluation at a negative timestep; this is an ad hoc choice outside the diffusion training range.
  • LAION-Art training subset = 0.2M of 8M
    The authors observed 0.2M real samples sufficed; the exact subset is unspecified, affecting reproduction.
assumptions (6)
  • ad hoc to paper Denoiser extrapolates to t=0 and t=-1
    Eq. (1) sets input timestep 0 and Eq. (4) uses timestep -1, both outside the DDPM noise schedule {1,...,T}; the paper only provides empirical evidence, no formal justification for this extrapolation.
  • domain assumption Blended latent preserves shared image-depth features
    Eq. (2) defines b0 = gamma x0 + (1-gamma) d0; the claim that this retains generative detail while adapting to depth is supported only by qualitative ablations.
  • domain assumption DINOv2/DAv2 predicts low-frequency depth accurately
    Section 3.3 relies on the observation that DINOv2-based models generalize but miss high-frequency details; this motivates Eq. (4)'s filter, but no direct measurement of frequency content is provided.
  • domain assumption MiDaS losses apply in latent space
    Eqs. (5)-(6) compute MAE and gradient matching on Stable Diffusion latents; the paper assumes latent-space errors align with depth-map quality.
  • domain assumption VAE reconstructs depth latents with negligible error
    Section 3.1 states the depth latent can be decoded with negligible error, relying on the SD VAE for a downstream task it was not trained for.
  • standard math Diffusion forward process and v-prediction are standard background
    Eqs. (2)-(3) follow Ho et al. and Salimans and Ho; these are accepted standard formulations.

how reviews work

0 comments
Cite this review

Pith. "Pith review of FiffDepth: Feed-forward Transformation of Diffusion-Based Generators for Detailed Depth Estimation." pith.science (2026). https://pith.science/paper/3GGOPQUA

@misc{pith2026241200671,
  author       = {Pith},
  title        = {Pith review of: FiffDepth: Feed-forward Transformation of Diffusion-Based Generators for Detailed Depth Estimation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/3GGOPQUA}},
  note         = {Machine review of arXiv:2412.00671}
}
read the original abstract

Monocular Depth Estimation (MDE) is a fundamental 3D vision problem with numerous applications such as 3D scene reconstruction, autonomous navigation, and AI content creation. However, robust and generalizable MDE remains challenging due to limited real-world labeled data and distribution gaps between synthetic datasets and real data. Existing methods often struggle with real-world test data with low efficiency, reduced accuracy, and lack of detail. To address these issues, we propose an efficient MDE approach named FiffDepth. The key feature of FiffDepth is its use of diffusion priors. It transforms diffusion-based image generators into a feed-forward architecture for detailed depth estimation. FiffDepth preserves key generative features and integrates the strong generalization capabilities of models like DINOv2. Through benchmark evaluations, we demonstrate that FiffDepth achieves exceptional accuracy, stability, and fine-grained detail, offering significant improvements in MDE performance against state-of-the-art MDE approaches. The paper's source code is available here: https://yunpeng1998.github.io/FiffDepth/

Figures

Figures reproduced from arXiv: 2412.00671 by the authors.

Figure 1
Figure 1. Compared to other methods, our model achieves more accurate details and better generalization in depth estimation. The final [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Overview of the proposed method. To simplify the representation, all the images we used above correspond to the respective latents. We transform the pre-trained diffusion model into a feed-forward approach for depth prediction, using only the result at t = 0 as the output during inference. During training, at t = 0, we use synthetic data to ensure detailed results, while at t = −1, we leverage pseudo-labels generate… view at source ↗
Figure 3
Figure 3. Filter learning. We use a learnable filter to map our results to detail levels similar to DINOv2’s, matching its outputs and thereby transferring DINOv2’s generalization capabilities to our model without compromising our inherent details. to supervise d−1, allowing us to transfer DINOv2’s robust￾ness without interfering with the detailed features in d0. In this process, we use real-world image data for x0 [PITH_FUL… view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: Qualitative comparison across different datasets. Our method is capable of predicting the depth of various fine objects, such as lampposts, railings, and chair legs [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]
Figure 5
Figure 5. Figure 5: Qualitative comparison on special scenarios. In the special scenarios of games, artworks, AI-generated content, and movies, our method demonstrates strong generalization capability and the ability to predict detailed depth [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]
Figure 6
Figure 6. Figure 6: Boundary visualization comparison. These samples are from the AM-2k dataset. Method Sintel F1↑ Spring F1↑ iBims F1↑ AM R↑ P3M R↑ DIS R↑ DepthAnything v2 0.228 0.056 0.111 0.107 0.131 0.056 Depth Pro 0.409 0.079 0.176 0.173 0.168 0.077 FiffDepth (Ours) 0.423 0.086 0.189…
Figure 7
Figure 7. Figure 7: Ablation studies. The generalization capability and depth details of the method are affected when some essential com￾ponents are missing. recent work Depth Pro [3] to evaluate boundary sharpness. Following Depth Pro, we compute the depth average bound￾ary F1 score for …

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

59 extracted references · 22 canonical work pages

  1. [1]

    Bidirectional attention network for monocular depth estimation

    Shubhra Aich, Jean Marie Uwabeza Vianney, Md Amirul Is- lam, and Mannat Kaur Bingbing Liu. Bidirectional attention network for monocular depth estimation. In 2021 IEEE In- ternational Conference on Robotics and Automation (ICRA), pages 11746–11752. IEEE, 2021. 2

    work page 2021

  2. [2]

    Generalized denoising auto-encoders as generative models

    Yoshua Bengio, Li Yao, Guillaume Alain, and Pascal Vin- cent. Generalized denoising auto-encoders as generative models. Advances in neural information processing systems, 26, 2013. 2

    work page 2013

  3. [3]

    Depth pro: Sharp monocular metric depth in less than a second

    Aleksei Bochkovskii, Ama ¨el Delaunoy, Hugo Germain, Marcel Santos, Yichao Zhou, Stephan R Richter, and Vladlen Koltun. Depth pro: Sharp monocular metric depth in less than a second. arXiv preprint arXiv:2410.02073, 2024. 8

    arXiv 2024

  4. [4]

    A naturalistic open source movie for opti- cal flow evaluation

    Daniel J Butler, Jonas Wulff, Garrett B Stanley, and Michael J Black. A naturalistic open source movie for opti- cal flow evaluation. In Computer Vision–ECCV 2012: 12th European Conference on Computer Vision, Florence, Italy, October 7-13, 2012, Proceedings, Part VI 12 , pages 611–

    2012

  5. [5]

    Vir- tual kitti 2

    Yohann Cabon, Naila Murray, and Martin Humenberger. Vir- tual kitti 2. arXiv preprint arXiv:2001.10773, 2020. 5

    arXiv 2001

  6. [6]

    Single- image depth perception in the wild

    Weifeng Chen, Zhao Fu, Dawei Yang, and Jia Deng. Single- image depth perception in the wild. Advances in neural in- formation processing systems, 29, 2016. 2

    work page 2016

  7. [7]

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

    Weifeng Chen, Shengyi Qian, David Fan, Noriyuki Ko- jima, Max Hamilton, and Jia Deng. Oasis: A large-scale dataset for single image 3d in the wild. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 679–688, 2020. 2

    work page 2020

  8. [8]

    Self-supervised learning with geometric constraints in monocular video: Connecting flow, depth, and camera

    Yuhua Chen, Cordelia Schmid, and Cristian Sminchis- escu. Self-supervised learning with geometric constraints in monocular video: Connecting flow, depth, and camera. In Proceedings of the IEEE/CVF international conference on computer vision, pages 7063–7072, 2019. 2

    2019

Show all 59 references
  1. [9]

    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 Proceedings of the IEEE conference on computer vision and pattern recognition, pages 5828–5839, 2017. 5

    2017

  2. [10]

    Omnidata: A scalable pipeline for making multi- task mid-level vision datasets from 3d scans

    Ainaz Eftekhar, Alexander Sax, Jitendra Malik, and Amir Zamir. Omnidata: A scalable pipeline for making multi- task mid-level vision datasets from 3d scans. In Proceedings of the IEEE/CVF International Conference on Computer Vi- sion, pages 10786–10796, 2021. 2, 5

    2021

  3. [11]

    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. Advances in neural information processing systems , 27, 2014. 2

    2014

  4. [12]

    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. In Proceedings of the IEEE conference on computer vision and pattern recogni- tion, pages 2002–2011, 2018. 2

    2002

  5. [13]

    Geowiz- ard: Unleashing the diffusion priors for 3d geometry estima- tion from a single image

    Xiao Fu, Wei Yin, Mu Hu, Kaixuan Wang, Yuexin Ma, Ping Tan, Shaojie Shen, Dahua Lin, and Xiaoxiao Long. Geowiz- ard: Unleashing the diffusion priors for 3d geometry estima- tion from a single image. arXiv preprint arXiv:2403.12013,

    arXiv

  6. [14]

    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 2012 IEEE conference on computer vision and pat- tern recognition, pages 3354–3361. IEEE, 2012. 5

    2012

  7. [15]

    Depthfm: Fast monocular depth estimation with flow matching

    Ming Gui, Johannes S Fischer, Ulrich Prestel, Pingchuan Ma, Dmytro Kotovenko, Olga Grebenkova, Stefan Andreas Baumann, Vincent Tao Hu, and Bj ¨orn Ommer. Depthfm: Fast monocular depth estimation with flow matching. arXiv preprint arXiv:2403.13788, 2024. 5

    2024 arXiv

  8. [16]

    Lotus: Diffusion-based visual foundation model for high-quality dense prediction

    Jing He, Haodong Li, Wei Yin, Yixun Liang, Leheng Li, Kaiqiang Zhou, Hongbo Liu, Bingbing Liu, and Ying- Cong Chen. Lotus: Diffusion-based visual foundation model for high-quality dense prediction. arXiv preprint arXiv:2409.18124, 2024. 3, 4, 5

    2024 arXiv

  9. [17]

    Masked autoencoders are scalable vision learners

    Kaiming He, Xinlei Chen, Saining Xie, Yanghao Li, Piotr Doll´ar, and Ross Girshick. Masked autoencoders are scalable vision learners. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages 16000– 16009, 2022. 2

    2022

  10. [18]

    Denoising dif- fusion probabilistic models

    Jonathan Ho, Ajay Jain, and Pieter Abbeel. Denoising dif- fusion probabilistic models. Advances in neural information processing systems, 33:6840–6851, 2020. 5

    2020

  11. [19]

    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 Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2024...

    2024

  12. [20]

    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. In Proceedings of the European Con- ference on Computer Vision (ECCV) Workshops, pages 0–0,

  13. [21]

    Ro- bust consistent video depth estimation

    Johannes Kopf, Xuejian Rong, and Jia-Bin Huang. Ro- bust consistent video depth estimation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 1611–1621, 2021. 2

    2021

  14. [22]

    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. 2

    1907 arXiv

  15. [23]

    Privacy- preserving portrait matting

    Jizhizi Li, Sihan Ma, Jing Zhang, and Dacheng Tao. Privacy- preserving portrait matting. In Proceedings of the 29th ACM international conference on multimedia , pages 3501–3509,

  16. [24]

    Bridging composite and real: towards end-to-end deep image matting

    Jizhizi Li, Jing Zhang, Stephen J Maybank, and Dacheng Tao. Bridging composite and real: towards end-to-end deep image matting. International Journal of Computer Vision , 130(2):246–266, 2022. 8

    2022

  17. [25]

    Megadepth: Learning single- view depth prediction from internet photos

    Zhengqi Li and Noah Snavely. Megadepth: Learning single- view depth prediction from internet photos. In Proceed- ings of the IEEE conference on computer vision and pattern recognition, pages 2041–2050, 2018. 2

    2018

  18. [26]

    Depthformer: Exploiting long-range correlation and local in- formation for accurate monocular depth estimation.Machine Intelligence Research, 20(6):837–854, 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, 20(6):837–854, 2023. 2

    2023

  19. [27]

    Fine-tuning image-conditional diffusion models is easier than you think

    Gonzalo Martin Garcia, Karim Abou Zeid, Christian Schmidt, Daan de Geus, Alexander Hermans, and Bastian Leibe. Fine-tuning image-conditional diffusion models is easier than you think. In Proceedings of the IEEE/CVF Win- ter Conference on Applications of Computer Vision (WACV),

  20. [28]

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

    Lukas Mehl, Jenny Schmalfuss, Azin Jahedi, Yaroslava Nali- vayko, and Andr ´es Bruhn. Spring: A high-resolution high- detail dataset and benchmark for scene flow, optical flow and stereo. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pa...

    2023

  21. [29]

    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, 4

    2023 arXiv

  22. [30]

    P3depth: Monocular depth estimation with a piecewise planarity prior

    Vaishakh Patil, Christos Sakaridis, Alexander Liniger, and Luc Van Gool. P3depth: Monocular depth estimation with a piecewise planarity prior. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages 1610–1621, 2022. 2

    2022

  23. [31]

    Highly accurate dichotomous im- age segmentation

    Xuebin Qin, Hang Dai, Xiaobin Hu, Deng-Ping Fan, Ling Shao, and Luc Van Gool. Highly accurate dichotomous im- age segmentation. In European Conference on Computer Vi- sion, pages 38–56. Springer, 2022. 8

    2022

  24. [32]

    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 transactions on pattern analysis and machine intelligence, 44(3):1623–1637, 2020. 1, 2, 5

    2020

  25. [33]

    Vi- sion transformers for dense prediction

    Ren ´e Ranftl, Alexey Bochkovskiy, and Vladlen Koltun. Vi- sion transformers for dense prediction. In Proceedings of the IEEE/CVF international conference on computer vision, pages 12179–12188, 2021. 1, 2, 5

    2021

  26. [34]

    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 Proceedings of the IEEE/CVF international conference o...

    2021

  27. [35]

    High-resolution image synthesis with latent diffusion models

    Robin Rombach, Andreas Blattmann, Dominik Lorenz, Patrick Esser, and Bj ¨orn Ommer. High-resolution image synthesis with latent diffusion models. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 10684–10695, 2022. 2, 3

    2022

  28. [36]

    Progressive distillation for fast sampling of diffusion models

    Tim Salimans and Jonathan Ho. Progressive distillation for fast sampling of diffusion models. arXiv preprint arXiv:2202.00512, 2022. 4

    2022 arXiv

  29. [37]

    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 Proceed- ings of the IEEE conference on computer vision and ...

    2017

  30. [38]

    Laion-5b: An open large-scale dataset for training next generation image-text models

    Christoph Schuhmann, Romain Beaumont, Richard Vencu, Cade Gordon, Ross Wightman, Mehdi Cherti, Theo Coombes, Aarush Katta, Clayton Mullis, Mitchell Worts- man, et al. Laion-5b: An open large-scale dataset for training next generation image-text models. Advances in Neural In- f...

    2022

  31. [39]

    Realmdreamer: Text-driven 3d scene gener- ation with inpainting and depth diffusion

    Jaidev Shriram, Alex Trevithick, Lingjie Liu, and Ravi Ra- mamoorthi. Realmdreamer: Text-driven 3d scene gener- ation with inpainting and depth diffusion. arXiv preprint arXiv:2404.07199, 2024. 1

    2024 arXiv

  32. [40]

    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 Computer Vision–ECCV 2012: 12th Eu- ropean Conference on Computer Vision, Florence, Italy, Oc- tober 7-13, 2012, Proceedings, Part V 12 , pages 746–760...

    2012

  33. [41]

    Deepv2d: Video to depth with differentiable structure from motion

    Zachary Teed and Jia Deng. Deepv2d: Video to depth with differentiable structure from motion. arXiv preprint arXiv:1812.04605, 2018. 2

    2018 arXiv

  34. [42]

    Diode: A dense indoor and outdoor depth dataset

    Igor Vasiljevic, Nick Kolkin, Shanyi Zhang, Ruotian Luo, Haochen Wang, Falcon Z Dai, Andrea F Daniele, Moham- madreza Mostajabi, Steven Basart, Matthew R Walter, et al. Diode: A dense indoor and outdoor depth dataset. arXiv preprint arXiv:1908.00463, 2019. 5

    1908 arXiv

  35. [43]

    Extracting and composing robust features with denoising autoencoders

    Pascal Vincent, Hugo Larochelle, Yoshua Bengio, and Pierre-Antoine Manzagol. Extracting and composing robust features with denoising autoencoders. In Proceedings of the 25th international conference on Machine learning , pages 1096–1103, 2008. 2

    2008

  36. [44]

    Sparsenerf: Distilling depth ranking for few-shot novel view synthesis

    Guangcong Wang, Zhaoxi Chen, Chen Change Loy, and Zi- wei Liu. Sparsenerf: Distilling depth ranking for few-shot novel view synthesis. InProceedings of the IEEE/CVF Inter- national Conference on Computer Vision, pages 9065–9076,

  37. [45]

    Pseudo- lidar from visual depth estimation: Bridging the gap in 3d object detection for autonomous driving

    Yan Wang, Wei-Lun Chao, Divyansh Garg, Bharath Hari- haran, Mark Campbell, and Kilian Q Weinberger. Pseudo- lidar from visual depth estimation: Bridging the gap in 3d object detection for autonomous driving. In Proceedings of the IEEE/CVF conference on computer vision and patt...

    2019

  38. [46]

    Neural video depth stabilizer

    Yiran Wang, Min Shi, Jiaqi Li, Zihao Huang, Zhiguo Cao, Jianming Zhang, Ke Xian, and Guosheng Lin. Neural video depth stabilizer. In Proceedings of the IEEE/CVF Interna- tional Conference on Computer Vision , pages 9466–9476,

  39. [47]

    Diffusion models trained with large data are transferable vi- sual models

    Guangkai Xu, Yongtao Ge, Mingyu Liu, Chengxiang Fan, Kangyang Xie, Zhiyue Zhao, Hao Chen, and Chunhua Shen. Diffusion models trained with large data are transferable vi- sual models. arXiv preprint arXiv:2403.06090, 2024. 3, 4, 5

    2024 arXiv

  40. [48]

    Transformer-based attention networks for continuous pixel-wise prediction

    Guanglei Yang, Hao Tang, Mingli Ding, Nicu Sebe, and Elisa Ricci. Transformer-based attention networks for continuous pixel-wise prediction. In Proceedings of the IEEE/CVF International Conference on Computer vision , pages 16269–16279, 2021. 2

    2021

  41. [49]

    Depth anything: Unleashing the power of large-scale unlabeled data

    Lihe Yang, Bingyi Kang, Zilong Huang, Xiaogang Xu, Jiashi Feng, and Hengshuang Zhao. Depth anything: Unleashing the power of large-scale unlabeled data. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 10371–10381, 2024. 2, 5

    2024

  42. [50]

    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, 5

    2024 arXiv

  43. [51]

    Diffusion model as repre- sentation learner

    Xingyi Yang and Xinchao Wang. Diffusion model as repre- sentation learner. In Proceedings of the IEEE/CVF Interna- tional Conference on Computer Vision, pages 18938–18949,

  44. [52]

    Mamo: Leveraging memory and attention for monocular video depth estimation

    Rajeev Yasarla, Hong Cai, Jisoo Jeong, Yunxiao Shi, Risheek Garrepalli, and Fatih Porikli. Mamo: Leveraging memory and attention for monocular video depth estimation. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 8754–8764, 2023. 2

    2023

  45. [53]

    Stablenormal: Reducing diffusion variance for stable and sharp normal

    Chongjie Ye, Lingteng Qiu, Xiaodong Gu, Qi Zuo, Yushuang Wu, Zilong Dong, Liefeng Bo, Yuliang Xiu, and Xiaoguang Han. Stablenormal: Reducing diffusion variance for stable and sharp normal. arXiv preprint arXiv:2406.16864, 2024. 3

    2024 arXiv

  46. [54]

    Di- versedepth: Affine-invariant depth prediction using diverse data

    Wei Yin, Xinlong Wang, Chunhua Shen, Yifan Liu, Zhi Tian, Songcen Xu, Changming Sun, and Dou Renyin. Di- versedepth: Affine-invariant depth prediction using diverse data. arXiv preprint arXiv:2002.00569, 2020. 2, 5

    2002 arXiv

  47. [55]

    Learning to recover 3d scene shape from a single image

    Wei Yin, Jianming Zhang, Oliver Wang, Simon Niklaus, Long Mai, Simon Chen, and Chunhua Shen. Learning to recover 3d scene shape from a single image. In Proceed- ings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 204–213, 2021

    2021

  48. [56]

    Hierarchical normalization for robust monocular depth estimation

    Chi Zhang, Wei Yin, Billzb Wang, Gang Yu, Bin Fu, and Chunhua Shen. Hierarchical normalization for robust monocular depth estimation. Advances in Neural Informa- tion Processing Systems, 35:14128–14139, 2022. 5

    2022

  49. [57]

    A tale of two features: Stable diffusion complements dino for zero-shot semantic correspondence

    Junyi Zhang, Charles Herrmann, Junhwa Hur, Luisa Pola- nia Cabrera, Varun Jampani, Deqing Sun, and Ming-Hsuan Yang. A tale of two features: Stable diffusion complements dino for zero-shot semantic correspondence. Advances in Neural Information Processing Systems, 36, 2024. 3

    2024

  50. [58]

    Betterdepth: Plug-and-play diffu- sion refiner for zero-shot monocular depth estimation

    Xiang Zhang, Bingxin Ke, Hayko Riemenschneider, Nando Metzger, Anton Obukhov, Markus Gross, Konrad Schindler, and Christopher Schroers. Betterdepth: Plug-and-play diffu- sion refiner for zero-shot monocular depth estimation. arXiv preprint arXiv:2407.17952, 2024. 3

    2024 arXiv

  51. [59]

    Unleashing text-to-image diffu- sion models for visual perception

    Wenliang Zhao, Yongming Rao, Zuyan Liu, Benlin Liu, Jie Zhou, and Jiwen Lu. Unleashing text-to-image diffu- sion models for visual perception. In Proceedings of the IEEE/CVF International Conference on Computer Vision , pages 5729–5739, 2023. 3

    2023

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.