Pith. sign in

REVIEW 4 major objections 5 minor 58 references

All-in-One: Transferring Vision Foundation Models into Stereo Matching

T0 review · 4 major / 5 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read AIO-Stereo claims that transferring selected knowledge from multiple vision foundation models into a CNN stereo network yields state-of-the-art disparity estimation, including first place on Middlebury and best published scores on ETH3D.

desk verdict Plausible method with real benchmark gains, but the zero-shot improvement is never isolated from the distillation loss that supposedly causes it. read the letter →

arxiv 2412.09912 v1 pith:BDZSOX3B submitted 2024-12-13 cs.CV

Jingyi Zhou , Haoyu Zhang , Jiakang Yuan , Peng Ye , Tao Chen , Hao Jiang , Meiya Chen , Yangyang Zhang This is my paper · ORCID
classification cs.CV
keywords stereomatchingvisionfoundationmodelsknowledgedistillationdual-levelselectivetransfermixtureofexpertsfeaturealignmentzero-shotgeneralizationdisparityestimation
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

AIO-Stereo aims to show that stereo matching improves when the feature encoder, not just the iterative refinement, is strengthened by knowledge from vision foundation models (VFMs). The paper proposes a framework that distills DINOv2, SAM, and Depth Anything v2 into a single CNN-based stereo network, aligning the heterogeneous feature spaces and letting a gating network choose the most useful expert features per pixel. The authors report that this selective transfer puts the method first on the Middlebury benchmark and ahead of all published methods on ETH3D, and that it improves zero-shot performance on unseen real data. If the claim holds, strong stereo accuracy can come from better encoded features rather than from redesigning the update module.

What carries the argument

The load-bearing component is the dual-level selective knowledge transfer (DLSKT) module. It operates at two levels: a distillation level, where an expert network plus a feature-alignment network maps each VFM's stage features into the CNN's latent space under an MSE loss $L_{\mathrm{KD},i}$; and a fusion level, where the gating network $g_i = \mathrm{KeepTopK}(\mathrm{Softmax}(G_i(f_i)), k)$ computes per-pixel weights and the selected expert features are added into the residual stream, $f_{i+1} = B_i(f_i) + \sum_{x \in \{d,s,a\}} e^x_i \odot g_i(x)$. This design is what lets heterogeneous Transformer-based VFM features be transferred into the CNN backbone while avoiding conflicts among conflicting knowledge sources.

What would settle it

A direct test would be to replace the learned gating with fixed uniform averaging of the three aligned expert features while keeping all other components identical. If the Middlebury and ETH3D gains persist, the selective mechanism is not responsible; if they collapse, the per-pixel selection carries the result. Alternatively, if zero-shot accuracy on a clearly new domain drops to the baseline, the transferred features are not general.

Watch

Extended reading notes

Core claim

The paper's central claim is that the quality of the encoded features matters as much as the iterative optimization in stereo matching, and that a single model can absorb complementary strengths of multiple vision foundation models. AIO-Stereo builds on the Selective-IGEV baseline and inserts a dual-level selective knowledge transfer module into the context network's residual blocks. Each of three VFMs (DINOv2 for robust foreground and semantic features, SAM for edges and small objects, Depth Anything v2 for dark and low-texture areas) has its own expert and feature-alignment network; one MSE distillation loss aligns the heterogeneous feature spaces, and a gating network with per-pixel top-k selection fuses only the most relevant experts. The resulting model ranks first on Middlebury, achieves the best published results on ETH3D on most metrics, and improves D1-fg on KITTI-2015, all without changing the iterative refinement machinery.

Load-bearing premise

The load-bearing premise is that the learned feature alignments and per-pixel gating genuinely transfer general visual knowledge from the three foundation models, rather than only fitting the specific training and validation benchmarks that produced the reported leaderboard numbers.

Editorial extensions

If this is right

  • Stereo matching can reach state-of-the-art accuracy by improving the encoder through VFM distillation, without redesigning the refinement module.
  • A single CNN backbone can simultaneously use several heterogeneous foundation models by aligning their features and per-pixel gating, resolving feature conflicts.
  • Zero-shot performance on unseen real datasets improves, indicating that the transferred features generalize beyond the training distribution.
  • The framework is flexible: adding or removing a VFM changes performance gracefully, so future foundation models can be plugged in without architectural changes.

Reading between the lines

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

  • The same selective multi-teacher transfer could be applied to other dense prediction tasks such as optical flow, monocular depth estimation, or semantic segmentation, where CNN encoders face similar feature-quality limits.
  • The per-pixel gating visualizations suggest a testable prediction: regions with dark or low texture should consistently route to Depth Anything, edges to SAM, and salient foregrounds to DINO; a quantitative region-wise study could confirm this division of labor.
  • As more specialized foundation models are released, the framework's benefit may grow, but the cost of running three VFM teachers during training, and the extra expert and alignment parameters, may become a bottleneck for real-time use.
  • Since the ablation validates the mechanism on a single training set, an independent multi-seed evaluation across several benchmarks would clarify how much of the gain is due to selection versus the added parameters.
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 / 5 minor

Summary. The paper proposes AIO-Stereo, a stereo matching method that transfers knowledge from three vision foundation models (DINOv2, SAM, Depth Anything v2) into a CNN-based iterative stereo network (built on Selective-IGEV). The method introduces expert networks and feature alignment networks for heterogeneous feature adaptation, a dual-level knowledge utilization that combines feature distillation with a residual forward-fusion connection, and a gating mechanism with KeepTopK selection to fuse expert features per pixel. The authors evaluate on Scene Flow, Middlebury, ETH3D, and KITTI-2015, reporting state-of-the-art or comparable results, including first place on the Middlebury leaderboard and best published results on ETH3D, plus a zero-shot Scene Flow-to-Middlebury experiment.

Significance. If the reported gains are robust, the paper makes a useful empirical contribution by showing that carefully aligned and selectively fused VFM features can improve a modern iterative stereo matcher, particularly in dark and low-texture regions. The design is novel in combining multi-teacher distillation with gated forward fusion for stereo matching, and the zero-shot evaluation is a positive feature. The strengths include external benchmark comparisons, a clear ablation intent, and qualitative visualizations of per-pixel expert selection. However, the central claims are currently not fully supported: the ablations are run on a single training set without statistical significance, the forward-fusion ablation is confounded by also dropping one teacher, the distillation loss is not gated despite the 'selective' claim, and the leaderboard-level SOTA statements go beyond the numbers actually shown. These issues affect the paper's main generalization and state-of-the-art claims, so the contribution cannot be accepted as-is.

major comments (4)
  1. [Table 1] The 'w/o Forward Fusion' ablation is confounded: this row removes the forward fusion path but also removes Depth Anything v2 from the model (the Depth Anything v2 column is unchecked). Consequently, the reported drop from 0.66/3.48 to 0.67/3.52 cannot be attributed solely to removing forward fusion; it may reflect the loss of a teacher. The authors should run a controlled ablation that removes forward fusion while keeping all three VFMs, and also remove Depth Anything v2 while keeping forward fusion, to isolate the two effects. As reported, the dual-level knowledge utilization claim is not cleanly supported.
  2. [Zero-Shot Generalization (Table 3)] The central generalization claim—that VFM distillation improves zero-shot performance—is never tested with the appropriate control. The ablation study in Table 1 is performed on the MiddEval v3 training set, where the model is fine-tuned on the target data, so improvements there could reflect fitting or regularization rather than transferable VFM knowledge. The zero-shot Scene Flow-to-Middlebury experiment in Table 3 only reports the full model. The authors should rerun the zero-shot protocol with the 'w/o Distillation' and 'w/o Selection' variants; if those variants match the full model under zero-shot transfer, the distillation gains observed on the training set do not support the claimed generalization benefit.
  3. [Eq. (6) and Eq. (8)] The 'selective' mechanism does not actually gate the distillation gradients. Eq. (6) sums the MSE distillation losses over DINO, SAM, and Depth Anything without any weighting by the gating logits computed in Eq. (7); the gating weights are applied only to the forward fusion path in Eq. (8). Therefore, conflicting or low-quality VFM knowledge can still propagate into the backbone during training through the distillation loss, undercutting the paper's stated motivation of preventing knowledge conflicts among VFMs. Either the distillation loss should be weighted by the gating probabilities, or the authors should explicitly justify why unweighted distillation does not reintroduce the conflicts the gating is designed to avoid.
  4. [Abstract and 'Comparisons with State-of-the-art'] The claims of ranking 1st on the Middlebury dataset and outperforming all published work on ETH3D are not substantiated by Table 2, which compares only a small hand-picked set of prior methods. No Middlebury or ETH3D leaderboard snapshot, date, or URL is provided, and 'all published work' is not a verifiable statement from the reported experiments. The authors should either provide the leaderboard evidence from the respective benchmark websites or soften the claims to 'state-of-the-art among the compared methods.'
minor comments (5)
  1. [Abstract] There is a typo: 'start-of-the-art' should be 'state-of-the-art.'
  2. [Comparisons with State-of-the-art] The text contains typographical errors such as 'finstune' instead of 'fine-tune' and 'Selevtive-IGEV' instead of 'Selective-IGEV.'
  3. [Eq. (7)] The hyperparameter k in KeepTopK is never specified in the implementation details, and no sensitivity analysis is provided for it; the authors should report the value used and, ideally, an ablation.
  4. [Implementation Details] The description of the feature alignment network as 'heavier' and initialized with a 'higher learning rate' is not quantified; the architecture sizes, learning rate multipliers, and decay factors should be listed.
  5. [Eq. (4)] The distillation loss uses MSE between aligned expert features and VFM features, but the authors do not discuss whether the VFM features are L2-normalized or otherwise scaled; this can strongly affect MSE-based distillation and should be clarified.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the method is a standard supervised distillation pipeline evaluated on external benchmarks; no equation or fitted parameter is relabeled as a prediction.

full rationale

The paper's derivation chain is self-contained against external evidence. The stereo backbone is trained with the standard prediction loss (Eq. 2) and an MSE distillation loss (Eq. 4) against three frozen, externally pretrained VFMs (DINOv2, SAM, Depth Anything). The expert, feature-alignment, and gating networks are learnable modules trained end-to-end with ordinary supervised losses; their parameters are fit to data rather than derived from the target result, and the paper never renames fitted parameters as predictions. The gating in Eq. (7)-(8) is a learned selection mechanism, not an assumed conclusion, and the ablation in Table 1 tests it on the MiddEval v3 training set. Tables 2-3 report results on independent Middlebury, ETH3D, KITTI, and zero-shot Scene Flow-to-Middlebury evaluations, so the SOTA and zero-shot claims are empirical claims against external benchmarks rather than consequences of the paper's definitions. There are no self-citations or imported uniqueness theorems; the citation to Selective-IGEV is only as a baseline, and the VFM citations are to external pretrained models. The only notable weakness is that the zero-shot distillation control (e.g., a w/o-Distillation variant under the Scene Flow-to-Middlebury protocol) is absent, so the attribution of zero-shot gains to VFM knowledge is not fully isolated; this is a missing control, not a circular derivation. No equation in the paper reduces by construction to its own input, so the circularity score is 0.

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

No new physical entities are introduced. The free parameters are the unspecified hyperparameters and network architecture details that affect the reported results. The axioms are the domain assumptions about VFM knowledge transfer and the empirical validation of the gating mechanism.

free parameters (4)
  • KeepTopK k = not reported
    Number of experts retained per pixel in gating equation (7); value affects behavior but is not given.
  • distillation decay factor gamma_KD = not reported
    Used in final loss Eq. (9); not specified.
  • feature alignment network learning rate multiplier = not reported
    The paper states a higher initial learning rate and larger decay factor for the alignment network but does not report exact values.
  • expert/alignment network sizes = not reported
    Only described as 'lightweight' and 'heavier'; no layer counts or channels given, affecting capacity and results.
assumptions (4)
  • domain assumption Knowledge from VFMs is transferable to stereo matching via feature alignment and distillation.
    The paper assumes the general representations from DINOv2, SAM, and Depth Anything v2 contain useful signal for disparity estimation, motivated by their strong generalization.
  • ad hoc to paper The gating network trained with task loss learns complementary selection of VFM knowledge.
    No theoretical guarantee or analysis; the selection is validated only empirically in ablations and visualization.
  • domain assumption The three chosen VFMs are sufficiently complementary and non-conflicting after alignment.
    The paper argues DINO focuses on foreground, SAM on edges, and Depth Anything on dark/textureless areas, and relies on this complementarity.
  • standard math Standard deep learning optimization assumptions (e.g., SGD convergence, MSE loss behavior).
    Implicit in all training-based claims.

how reviews work

0 comments
Cite this review

Pith. "Pith review of All-in-One: Transferring Vision Foundation Models into Stereo Matching." pith.science (2026). https://pith.science/paper/BDZSOX3B

@misc{pith2026241209912,
  author       = {Pith},
  title        = {Pith review of: All-in-One: Transferring Vision Foundation Models into Stereo Matching},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/BDZSOX3B}},
  note         = {Machine review of arXiv:2412.09912}
}
abstract

As a fundamental vision task, stereo matching has made remarkable progress. While recent iterative optimization-based methods have achieved promising performance, their feature extraction capabilities still have room for improvement. Inspired by the ability of vision foundation models (VFMs) to extract general representations, in this work, we propose AIO-Stereo which can flexibly select and transfer knowledge from multiple heterogeneous VFMs to a single stereo matching model. To better reconcile features between heterogeneous VFMs and the stereo matching model and fully exploit prior knowledge from VFMs, we proposed a dual-level feature utilization mechanism that aligns heterogeneous features and transfers multi-level knowledge. Based on the mechanism, a dual-level selective knowledge transfer module is designed to selectively transfer knowledge and integrate the advantages of multiple VFMs. Experimental results show that AIO-Stereo achieves start-of-the-art performance on multiple datasets and ranks $1^{st}$ on the Middlebury dataset and outperforms all the published work on the ETH3D benchmark.

Figures

Figures reproduced from arXiv: 2412.09912 by the authors.

Figure 1
Figure 1. (a) The overview of AIO-Stereo which transfers [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. The Overall framework of AIO-Stereo. Left: AIO-Stereo selectively learns knowledge from SAM, DINO and Depth [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 1
Figure 1. As a result, indiscriminate acceptance of knowledge [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Visualization of the selection weights for each [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]
Figure 3
Figure 3. Figure 3: Visual comparison on the Middlebury dataset. [PITH_FULL_IMAGE:figures/full_fig_p007_3.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

58 extracted references · 34 canonical work pages

  1. [1]

    Bai, H.; Wu, J.; King, I.; and Lyu, M. 2020. Few shot network compression via cross distillation. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 34, 3203--3210

    work page 2020

  2. [2]

    Bao, W.; Wang, W.; Xu, Y.; Guo, Y.; Hong, S.; and Zhang, X. 2020. Instereo2k: a large real dataset for stereo matching in indoor scenes. Science China Information Sciences, 63: 1--11

    work page 2020

  3. [3]

    Boykov, Y.; Veksler, O.; and Zabih, R. 2001. Fast approximate energy minimization via graph cuts. IEEE Transactions on pattern analysis and machine intelligence, 23(11): 1222--1239

    work page 2001

  4. [4]

    J.; Wulff, J.; Stanley, G

    Butler, D. J.; Wulff, J.; Stanley, G. B.; and Black, M. J. 2012. A naturalistic open source movie for optical flow evaluation. In Computer Vision--ECCV 2012: 12th European Conference on Computer Vision, Florence, Italy, October 7-13, 2012, Proceedings, Part VI 12, 611--625. Springer

    2012

  5. [5]

    Caron, M.; Touvron, H.; Misra, I.; J\'egou, H.; Mairal, J.; Bojanowski, P.; and Joulin, A. 2021. Emerging Properties in Self-Supervised Vision Transformers. In Proceedings of the International Conference on Computer Vision (ICCV)

    2021

  6. [6]

    Chang, J.-R.; and Chen, Y.-S. 2018 a . Pyramid stereo matching network. In CVPR, 5410--5418

    work page 2018

  7. [7]

    Chang, J.-R.; and Chen, Y.-S. 2018 b . Pyramid stereo matching network. In Proceedings of the IEEE conference on computer vision and pattern recognition, 5410--5418

    work page 2018

  8. [8]

    Cheng, X.; Zhong, Y.; Harandi, M.; Dai, Y.; Chang, X.; Li, H.; Drummond, T.; and Ge, Z. 2020. Hierarchical neural architecture search for deep stereo matching. NeurIPS, 33: 22158--22169

    work page 2020

Show all 58 references
  1. [9]

    Darcet, T.; Oquab, M.; Mairal, J.; and Bojanowski, P. 2023. Vision Transformers Need Registers

    2023

  2. [10]

    Dosovitskiy, A.; Beyer, L.; Kolesnikov, A.; Weissenborn, D.; Zhai, X.; Unterthiner, T.; Dehghani, M.; Minderer, M.; Heigold, G.; Gelly, S.; et al. 2020. An image is worth 16x16 words: Transformers for image recognition at scale. arXiv preprint arXiv:2010.11929

    2020 arXiv

  3. [11]

    Fukuda, T.; Suzuki, M.; Kurata, G.; Thomas, S.; Cui, J.; and Ramabhadran, B. 2017. Efficient Knowledge Distillation from an Ensemble of Teachers. In Interspeech, 3697--3701

    2017

  4. [12]

    Geiger, A.; Lenz, P.; and Urtasun, R. 2012. Are we ready for autonomous driving? the kitti vision benchmark suite. In CVPR, 3354--3361. IEEE

    2012

  5. [13]

    Hao, Z.; Guo, J.; Han, K.; Tang, Y.; Hu, H.; Wang, Y.; and Xu, C. 2024. One-for-All: Bridge the Gap Between Heterogeneous Architectures in Knowledge Distillation. Advances in Neural Information Processing Systems, 36

    2024

  6. [14]

    Hinton, G.; Vinyals, O.; and Dean, J. 2015. Distilling the knowledge in a neural network. arXiv preprint arXiv:1503.02531

    2015 arXiv

  7. [15]

    Hirschmuller, H. 2005. Accurate and efficient stereo processing by semi-global matching and mutual information. In 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'05), volume 2, 807--814. IEEE

    2005

  8. [16]

    R.; and Garibaldi, J

    Hirschm \"u ller, H.; Innocent, P. R.; and Garibaldi, J. 2002. Real-time correlation-based stereo vision with reduced border errors. International Journal of Computer Vision, 47: 229--246

    2002

  9. [17]

    A.; Jordan, M

    Jacobs, R. A.; Jordan, M. I.; Nowlan, S. J.; and Hinton, G. E. 1991. Adaptive Mixtures of Local Experts. Neural Computation, 3(1): 79--87

    1991

  10. [18]

    Kendall, A.; Martirosyan, H.; Dasgupta, S.; Henry, P.; Kennedy, R.; Bachrach, A.; and Bry, A. 2017. End-to-end learning of geometry and context for deep stereo regression. In ICCV, 66--75

    2017

  11. [19]

    Kim, J.; Park, S.; and Kwak, N. 2018. Paraphrasing complex network: Network compression via factor transfer. Advances in neural information processing systems, 31

    2018

  12. [20]

    C.; Lo, W.-Y.; et al

    Kirillov, A.; Mintun, E.; Ravi, N.; Mao, H.; Rolland, C.; Gustafson, L.; Xiao, T.; Whitehead, S.; Berg, A. C.; Lo, W.-Y.; et al. 2023. Segment anything. arXiv preprint arXiv:2304.02643

    2023 arXiv

  13. [21]

    Klaus, A.; Sormann, M.; and Karner, K. 2006. Segment-based stereo matching using belief propagation and a self-adapting dissimilarity measure. In 18th International Conference on Pattern Recognition (ICPR'06), volume 3, 15--18. IEEE

    2006

  14. [22]

    Komodakis, N.; and Zagoruyko, S. 2017. Paying more attention to attention: improving the performance of convolutional neural networks via attention transfer. In ICLR

    2017

  15. [23]

    Li, J.; Wang, P.; Xiong, P.; Cai, T.; Yan, Z.; Yang, L.; Liu, J.; Fan, H.; and Liu, S. 2022. Practical stereo matching via cascaded recurrent network with adaptive correlation. In CVPR, 16263--16272

    2022

  16. [24]

    Lipson, L.; Teed, Z.; and Deng, J. 2021. Raft-stereo: Multilevel recurrent field transforms for stereo matching. In 3DV, 218--227. IEEE

    2021

  17. [25]

    Liu, I.-J.; Peng, J.; and Schwing, A. G. 2019. Knowledge flow: Improve upon your teachers. arXiv preprint arXiv:1904.05878

    2019 arXiv

  18. [26]

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

    2021

  19. [27]

    Mayer, N.; Ilg, E.; Hausser, P.; Fischer, P.; Cremers, D.; Dosovitskiy, A.; and Brox, T. 2016. A large dataset to train convolutional networks for disparity, optical flow, and scene flow estimation. In Proceedings of the IEEE conference on computer vision and pattern recogniti...

    2016

  20. [28]

    Mehak, M.; and Balasubramanian, V. N. 2018. Knowledge distillation from multiple teachers using visual explanations. Ph.D. thesis, Indian Institute of Technology Hyderabad

    2018

  21. [29]

    Menze, M.; and Geiger, A. 2015. Object scene flow for autonomous vehicles. In CVPR, 3061--3070

    2015

  22. [30]

    Oquab, M.; Darcet, T.; Moutakanni, T.; Vo, H. V.; Szafraniec, M.; Khalidov, V.; Fernandez, P.; Haziza, D.; Massa, F.; El-Nouby, A.; Howes, R.; Huang, P.-Y.; Xu, H.; Sharma, V.; Li, S.-W.; Galuba, W.; Rabbat, M.; Assran, M.; Ballas, N.; Synnaeve, G.; Misra, I.; Jegou, H.; Maira...

    2023

  23. [31]

    Park, S.; and Kwak, N. 2019. Feed: Feature-level ensemble for knowledge distillation. arXiv preprint arXiv:1909.10754

    2019 arXiv

  24. [32]

    W.; Hallacy, C.; Ramesh, A.; Goh, G.; Agarwal, S.; Sastry, G.; Askell, A.; Mishkin, P.; Clark, J.; et al

    Radford, A.; Kim, J. W.; Hallacy, C.; Ramesh, A.; Goh, G.; Agarwal, S.; Sastry, G.; Askell, A.; Mishkin, P.; Clark, J.; et al. 2021. Learning transferable visual models from natural language supervision. In International conference on machine learning, 8748--8763. PMLR

    2021

  25. [33]

    Scharstein, D.; Hirschm \"u ller, H.; Kitajima, Y.; Krathwohl, G.; Ne s i \'c , N.; Wang, X.; and Westling, P. 2014. High-resolution stereo datasets with subpixel-accurate ground truth. In GCPR, 31--42. Springer

    2014

  26. [34]

    L.; Galliani, S.; Sattler, T.; Schindler, K.; Pollefeys, M.; and Geiger, A

    Schops, T.; Schonberger, J. L.; Galliani, S.; Sattler, T.; Schindler, K.; Pollefeys, M.; and Geiger, A. 2017. A multi-view stereo benchmark with high-resolution images and multi-camera videos. In CVPR, 3260--3269

    2017

  27. [35]

    Shen, C.; Xue, M.; Wang, X.; Song, J.; Sun, L.; and Song, M. 2019. Customizing student networks from heterogeneous teachers via adaptive knowledge amalgamation. In Proceedings of the IEEE/CVF International Conference on Computer Vision, 3504--3513

    2019

  28. [36]

    Sun, J.; Zheng, N.-N.; and Shum, H.-Y. 2003. Stereo matching using belief propagation. IEEE Transactions on pattern analysis and machine intelligence, 25(7): 787--800

    2003

  29. [37]

    Tankovich, V.; Hane, C.; Zhang, Y.; Kowdle, A.; Fanello, S.; and Bouaziz, S. 2021 a . Hitnet: Hierarchical iterative tile refinement network for real-time stereo matching. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 14362--14372

    2021

  30. [38]

    Tankovich, V.; Hane, C.; Zhang, Y.; Kowdle, A.; Fanello, S.; and Bouaziz, S. 2021 b . Hitnet: Hierarchical iterative tile refinement network for real-time stereo matching. In CVPR, 14362--14372

    2021

  31. [39]

    Teed, Z.; and Deng, J. 2020. Raft: Recurrent all-pairs field transforms for optical flow. In ECCV, 402--419. Springer

    2020

  32. [40]

    Touvron, H.; Cord, M.; Douze, M.; Massa, F.; Sablayrolles, A.; and J \'e gou, H. 2021. Training data-efficient image transformers & distillation through attention. In International conference on machine learning, 10347--10357. PMLR

    2021

  33. [41]

    Tremblay, J.; To, T.; and Birchfield, S. 2018. Falling things: A synthetic dataset for 3d object detection and pose estimation. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition Workshops, 2038--2041

    2018

  34. [42]

    Van Meerbergen, G.; Vergauwen, M.; Pollefeys, M.; and Van Gool, L. 2002. A hierarchical symmetric stereo algorithm using dynamic programming. International Journal of Computer Vision, 47: 275--285

    2002

  35. [43]

    Wang, W.; Zhu, D.; Wang, X.; Hu, Y.; Qiu, Y.; Wang, C.; Hu, Y.; Kapoor, A.; and Scherer, S. 2020. Tartanair: A dataset to push the limits of visual slam. In 2020 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 4909--4916. IEEE

    2020

  36. [44]

    Wang, X.; Xu, G.; Jia, H.; and Yang, X. 2024. Selective-stereo: Adaptive frequency information selection for stereo matching. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 19701--19710

    2024

  37. [45]

    Wang, X.; Zhang, X.; Cao, Y.; Wang, W.; Shen, C.; and Huang, T. 2023. Seggpt: Segmenting everything in context. arXiv preprint arXiv:2304.03284

    2023 arXiv

  38. [46]

    Xu, G.; Wang, X.; Ding, X.; and Yang, X. 2023 a . Iterative Geometry Encoding Volume for Stereo Matching. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 21919--21928

    2023

  39. [47]

    Xu, G.; Wang, X.; Ding, X.; and Yang, X. 2023 b . Iterative Geometry Encoding Volume for Stereo Matching. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 21919--21928

    2023

  40. [48]

    Xu, H.; and Zhang, J. 2020. Aanet: Adaptive aggregation network for efficient stereo matching. In CVPR, 1959--1968

    2020

  41. [49]

    Xu, H.; Zhang, J.; Cai, J.; Rezatofighi, H.; Yu, F.; Tao, D.; and Geiger, A. 2023 c . Unifying flow, stereo and depth estimation. IEEE Transactions on Pattern Analysis and Machine Intelligence

    2023

  42. [50]

    Yang, G.; Manela, J.; Happold, M.; and Ramanan, D. 2019. Hierarchical deep stereo matching on high-resolution images. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 5515--5524

    2019

  43. [51]

    Yang, L.; Kang, B.; Huang, Z.; Xu, X.; Feng, J.; and Zhao, H. 2024 a . Depth Anything: Unleashing the Power of Large-Scale Unlabeled Data. In CVPR

    2024

  44. [52]

    Yang, L.; Kang, B.; Huang, Z.; Zhao, Z.; Xu, X.; Feng, J.; and Zhao, H. 2024 b . Depth Anything V2. arXiv:2406.09414

    2024 arXiv

  45. [53]

    Yang, Q.; Wang, L.; Yang, R.; Stew \'e nius, H.; and Nist \'e r, D. 2008. Stereo matching with color-weighted correlation, hierarchical belief propagation, and occlusion handling. IEEE transactions on pattern analysis and machine intelligence, 31(3): 492--504

    2008

  46. [54]

    Zbontar, J.; and LeCun, Y. 2015. Computing the stereo matching cost with a convolutional neural network. In Proceedings of the IEEE conference on computer vision and pattern recognition, 1592--1599

    2015

  47. [55]

    Zhao, H.; Zhou, H.; Zhang, Y.; Chen, J.; Yang, Y.; and Zhao, Y. 2023. High-Frequency Stereo Matching Network. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 1327--1336

    2023

  48. [56]

    Zhao, H.; Zhou, H.; Zhang, Y.; Zhao, Y.; Yang, Y.; and Ouyang, T. 2022. EAI-stereo: Error aware iterative network for stereo matching. In Proceedings of the Asian Conference on Computer Vision, 315--332

    2022

  49. [57]

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

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

  50. [58]

    write newline

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

Pith tools

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