Pith. sign in

REVIEW 3 major objections 6 minor 36 references

Estimation of Absolute Scale in Monocular SLAM Using Synthetic Data

T0 review · 3 major / 6 minor · reviewed 2026-08-14 · deepseek-v4-flash

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

Pith's one-line read Monocular SLAM's missing absolute scale can be estimated by a network trained on synthetic driving data alone, reaching accuracy comparable to real-data training.

desk verdict A solid, incremental improvement to monocular scale estimation, with a useful synthetic-data result for CNNs but an overstated abstract that ignores the LSTM transfer failure. read the letter →

arxiv 1909.00713 v1 pith:L7OXX7ZJ submitted 2019-09-02 cs.CV cs.RO

Danila Rukhovich , Daniel Mouritzen , Ralf Kaestner , Martin Rufli , Alexander Velizhev This is my paper · ORCID
classification cs.CVcs.RO
keywords monocularSLAMscaleestimationabsolutesyntheticdataCARLAKITTICNNbidirectionalLSTM
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

This paper argues that the missing absolute scale in monocular SLAM can be recovered by a neural network that regresses the distance between camera centers of consecutive frames, and that such a network can be trained entirely on synthetic simulator images. If true, this removes the need for costly real-data collection with ground-truth trajectories, since annotated data can be generated automatically from a driving simulator. The paper reports that a CNN trained only on CARLA synthetic images reaches a standard deviation of 0.105–0.121 m on real KITTI test sequences, close to the 0.092–0.113 m of the same network trained on KITTI. A bidirectional LSTM trained on KITTI does better still (0.064–0.083 m), confirming that temporal smoothing helps when real dynamics are available. The authors conclude that synthetic data is a viable training source and that visual photorealism matters less than trajectory and scene diversity.

What carries the argument

The central object is a regression network that takes a pair of consecutive 280x120 RGB images (concatenated into a 6-channel input), normalizes them to a fixed intrinsic camera model (focal length 250 px, principal point at 140,60), and outputs the absolute distance between the camera centers. The CNN has five convolutional layers with ELU activations and dropout; a variant appends a bidirectional many-to-one LSTM operating on the sequence of frame-pair embeddings, enabling the model to learn temporal priors about vehicle motion. The machinery carries the argument because the fixed normalization makes estimates invariant to different cameras, the recurrent layer exploits dynamics, and the synthetic-data pipeline supplies automatically annotated low-resolution pairs with varied weather, maps, and camera configurations.

What would settle it

Train a bidirectional LSTM on CARLA trajectories with artificially injected real-world noise matched to KITTI inter-frame distance statistics, then evaluate on the same KITTI test sequences; if the accuracy still does not reach the KITTI-trained LSTM level, the synthetic-to-real gap is not explained by trajectory smoothness alone.

Watch

Extended reading notes

Core claim

The central claim is that absolute scale in monocular SLAM can be estimated directly from low-resolution image pairs—without scene assumptions, known object sizes, or extra sensors—by training a CNN to regress the translation distance between camera centers. The paper's key discovery is that this scale estimator does not need real images with ground truth: a CNN trained only on 800K image pairs generated by the CARLA simulator generalizes to unseen real KITTI sequences almost as well as the same CNN trained on KITTI, with error standard deviations of 0.105–0.121 m versus 0.092–0.113 m. Adding a bidirectional LSTM over a window of 19 pairs further reduces the error to 0.064–0.083 m when trained on real KITTI data, because it learns smooth vehicle dynamics, but the same LSTM trained only on CARLA does not improve over the CNN, which the authors attribute to synthetic trajectories being smoother than real ones. The paper also reports that two unsupervised domain-adaptation methods produce no gain, suggesting that photorealism is not the bottleneck.

Load-bearing premise

The claim that synthetic training reaches real-data accuracy depends on simulated camera trajectories being representative of real vehicle motion; this is true for the feed-forward CNN but appears to fail for the LSTM, whose CARLA-trained version is noticeably less accurate.

Editorial extensions

If this is right

  • Monocular SLAM systems can recover metric scale without IMU, stereo, or depth sensors, using a lightweight network on 0.03 MP images.
  • Scale estimates improve classical bundle adjustment by constraining inter-frame distances, reducing scale drift and helping loop closure.
  • Because synthetic data works, sensor reconfiguration (new camera position or intrinsics) can be handled by generating new simulator data instead of collecting and labeling new real data.
  • Data diversity—more virtual maps with different road layouts—matters more than photorealism; domain adaptation did not improve accuracy.
  • A bidirectional LSTM that learns vehicle dynamics lowers scale-error standard deviation from about 0.18 m baseline to 0.064–0.083 m on KITTI test sequences.

Reading between the lines

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

  • We infer that injecting real trajectory statistics into a simulator could close the remaining LSTM gap, since the paper isolates the smoothness of synthetic trajectories as the cause of the degradation.
  • We infer the method may transfer to non-driving platforms only if the training trajectories are regenerated for those dynamics, because the recurrent layer is the component most tied to vehicle motion.
  • The near-zero benefit from domain adaptation suggests that scale estimation from image pairs relies on geometric displacement cues that survive domain shift, so future work could focus on trajectory diversity rather than image realism.
  • We infer that evaluating the estimated inter-frame distances inside a full SLAM trajectory (for example, with absolute trajectory error) would show how much the improved scale accuracy translates to overall map quality, which the paper does not directly measure.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 6 minor

Summary. The manuscript proposes deep network architectures for estimating absolute inter-frame camera distance in monocular SLAM, extending the baseline of Frost et al. (2017). It introduces a deeper CNN and unidirectional/bidirectional LSTM variants, and trains them on KITTI and/or synthetic CARLA data, with evaluation on KITTI sequences 00, 02, and 08. Table 2 reports reductions in the standard deviation of distance error from roughly 0.15-0.18 m for the baseline to 0.064-0.083 m for the best bidirectional LSTM (length 19) trained on KITTI, and about 0.105-0.121 m for a CNN trained only on CARLA. The paper also evaluates LSTM sequence length, synthetic map diversity, and domain adaptation with T2Net and CyCADA. The central claim is that synthetic-only training achieves scale-estimation accuracy similar to real-data training, and that visual photorealism is not important for this task.

Significance. If the claims hold, the work provides a practical low-resolution scale estimator that can be integrated into monocular SLAM without requiring real-world data annotation. The evaluation uses external ground-truth camera poses from KITTI and CARLA, so the results are not self-referential, and the ablations cover architectures, synthetic diversity, and domain adaptation. The reported accuracy improvement over the baseline is visible and consistent across sequences. However, the central synthetic-parity claim is currently broader than the evidence: it is directly supported only for the feed-forward CNN, not for the recurrent model that gives the best absolute accuracy, and the main hyperparameter is selected on the test set. These issues are fixable with additional experiments and revised claims, but they are load-bearing for the paper's headline result.

major comments (3)
  1. [Abstract, Conclusion, Section 4.3, Table 2] The statement that 'using only synthetic training inputs, we can achieve similar scale estimation accuracy as that obtained from real data' is not supported for the best-performing architecture. Comparing row 5 (CNN, KITTI-only, sigma = 0.092-0.113 m) with row 6 (CNN, CARLA-only, sigma = 0.105-0.121 m) supports the claim for the CNN, but rows 8 and 9 show that the bidirectional LSTM (B,19) degrades from sigma = 0.064-0.083 m with KITTI-only training to sigma = 0.102-0.132 m with CARLA-only training. The authors acknowledge in Section 4.3 that 'our virtual vehicle dynamics do not generalize well to the real dynamics,' yet the abstract and conclusion state the claim without this qualification. The claim should be explicitly scoped to feed-forward architectures, or the authors should provide a recurrent configuration whose synthetic-only training reaches KITTI-trained accuracy.
  2. [Section 4.4, Table 4, Table 2 rows 8-10] The LSTM sequence length and directionality are selected by sweeping values on the same KITTI test sequences (00, 02, 08) that are used to report the final results in Table 2. In particular, length 19 is chosen because Table 4 shows the best sigma for the bidirectional LSTM on the test set, and this exact configuration is then presented as 'our best results' in Table 2 rows 8-10. This is test-set selection and can inflate the reported gains. The authors should use a validation split drawn from the training sequences for hyperparameter selection, or report the entire sweep as exploratory and avoid singling out the best test-set configuration as the headline result.
  3. [Section 4.3, Tables 2-6] No error bars or multiple-run statistics are provided for any experiment. Several conclusions rest on small differences: rows 5 and 6 of Table 2 differ by 0.001-0.029 m depending on the sequence, and the domain-adaptation results in Table 6 differ by only 0.002-0.005 m. Without repeated training runs or another variance estimate, the claims of 'similar accuracy' and 'almost no improvement or even minor degradation' are not statistically grounded. The authors should report mean and standard deviation over at least three runs for the key comparisons, or explicitly justify why single runs suffice for the conclusions drawn.
minor comments (6)
  1. [Section 3.2] There is a typo: 'vehicle dymanics' should be 'vehicle dynamics.'
  2. [Table 4 caption] The caption spells 'arhitectures'; this should be 'architectures.'
  3. [References [26] and [27]] References [26] and [27] list the same paper by Scaramuzza et al. with different page ranges; these should be consolidated into one reference.
  4. [Section 4.5, Table 5] Table 5 reports performance of the CNN architecture trained on synthetic data, but the text should state explicitly that all rows use CARLA-only training and the same evaluation protocol as Table 2; currently this is implied rather than stated.
  5. [Figure 3 caption] The caption says the CNN block corresponds to the green block of Figure 2, but Figure 2's caption does not mention a green block; please make the color coding explicit in both captions.
  6. [Section 4.2, Table 3] Section 4.2 reports approximately 26K testing image pairs for sequences 00, 02, and 08, while Table 3 mentions 20K image pairs for 'all available KITTI sequences'; clarify which set of sequences and how many pairs are used in Table 3.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: central results are evaluated against external KITTI/CARLA ground truth.

full rationale

The paper contains no circular derivation. Scale estimates are compared against external ground-truth camera poses from KITTI and CARLA (Sections 4.2 and 4.3); no quantity used as a prediction is defined in terms of the fitted outputs. The CNN and LSTM are trained with MSE on a distance regression task and evaluated on held-out KITTI sequences (Table 2), and synthetic-data generalization is tested by training on CARLA-only data and evaluating on unseen real KITTI frames. The only internal model choice is the LSTM sequence length reported in Table 4; although that length is evaluated on the same KITTI test sequences, the reported sigma values are computed from external ground-truth distances, and no equation or fitted parameter is renamed as a prediction. The paper cites prior work [10] as a baseline and external domain-adaptation methods [1,34], but none of the load-bearing claims rests on a self-citation or on a uniqueness/ansatz assertion from the authors' own prior work. The abstract's unqualified 'similar scale accuracy' claim is broader than the evidence for the recurrent variant (rows 9 and 10 of Table 2), but that is a scope/correctness concern, not circularity.

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

The method has no mathematical derivation; it is an empirical regression pipeline. The central accuracy numbers depend on the ground-truth labels, the normalization choices, and the training-set composition. No nonstandard entities are introduced. The key hidden assumptions are that synthetic visual appearance and synthetic dynamics transfer to real data, and the paper's own tables show this transfer is architecture-dependent.

free parameters (4)
  • Dmax = 1.7 m
    Maximum camera displacement allowed for non-consecutive training pairs; chosen to match the test set with a 20 cm margin (Section 4.2).
  • LSTM sequence length = 19 (selected on test set)
    Section 4.4 evaluates lengths 5, 11, and 19 on KITTI test sequences; the headline result uses length 19, making the reported best number partly selected on the evaluation data.
  • Image normalization intrinsics = focal length 250 px, principal point (140, 60)
    The authors choose these replacement intrinsics in Section 4.2 to use the full image; the method depends on this normalization for invariance across cameras.
  • Training set scale = 800K synthetic pairs vs. 100K real pairs
    Synthetic and real training sets differ in size, confounding the synthetic-versus-real accuracy comparison; this is a design choice rather than a fitted number.
assumptions (4)
  • domain assumption Ground-truth camera trajectories for KITTI and CARLA are accurate enough to serve as supervision.
    All reported errors are differences from these labels; if the labels are biased, the sigma values shift. Section 4.2 describes deriving ground truth from KITTI and CARLA poses.
  • domain assumption The visual-geometric relationship between image pairs determines camera displacement independently of scene category and visual appearance.
    The method is a supervised black-box regressor, and transfer from synthetic to real images assumes this invariance. Section 4.3 rows 5 and 6 test it.
  • domain assumption Simulated trajectory dynamics in CARLA are representative enough for the recurrent model to learn useful temporal priors.
    Rows 9 and 10 of Table 2 show the LSTM trained on CARLA does not match the real-trained LSTM, contradicting this assumption for the LSTM variant.
  • domain assumption The image normalization using fixed intrinsics makes predictions invariant to different cameras.
    Section 4.2 changes focal length and principal point; the authors rely on this to apply one model across different camera configurations.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Estimation of Absolute Scale in Monocular SLAM Using Synthetic Data." pith.science (2026). https://pith.science/paper/L7OXX7ZJ

@misc{pith2026190900713,
  author       = {Pith},
  title        = {Pith review of: Estimation of Absolute Scale in Monocular SLAM Using Synthetic Data},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/L7OXX7ZJ}},
  note         = {Machine review of arXiv:1909.00713}
}
read the original abstract

This paper addresses the problem of scale estimation in monocular SLAM by estimating absolute distances between camera centers of consecutive image frames. These estimates would improve the overall performance of classical (not deep) SLAM systems and allow metric feature locations to be recovered from a single monocular camera. We propose several network architectures that lead to an improvement of scale estimation accuracy over the state of the art. In addition, we exploit a possibility to train the neural network only with synthetic data derived from a computer graphics simulator. Our key insight is that, using only synthetic training inputs, we can achieve similar scale estimation accuracy as that obtained from real data. This fact indicates that fully annotated simulated data is a viable alternative to existing deep-learning-based SLAM systems trained on real (unlabeled) data. Our experiments with unsupervised domain adaptation also show that the difference in visual appearance between simulated and real data does not affect scale estimation results. Our method operates with low-resolution images (0.03MP), which makes it practical for real-time SLAM applications with a monocular camera.

Figures

Figures reproduced from arXiv: 1909.00713 by the authors.

Figure 1
Figure 1. Example of real images from KITTI [12] (top row) and [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Baseline CNN architecture used in all experiments of this paper. [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Proposed bidirectional LSTM architecture. The CNN [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: Trajectories colored by the absolute distance estimation errors (KITTI sequences 00, 02, 08). Left: State-of-the-art results of [10] [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]
Figure 6
Figure 6. Figure 6: Histogram of distance estimation errors for LSTM (B, [PITH_FULL_IMAGE:figures/full_fig_p008_6.png]
Figure 7
Figure 7. Figure 7: Images with largest errors for LSTM (B, length 19). [PITH_FULL_IMAGE:figures/full_fig_p008_7.png]
Figure 5
Figure 5. Figure 5: Comparison of recovered distances (blue) and ground [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

36 extracted references · 32 canonical work pages

  1. [1]

    In International Conference on Machine Learning (ICML) , pages 1994–2003, 2018

    Cycada: Cycle consistent adversarial domain adaptation. In International Conference on Machine Learning (ICML) , pages 1994–2003, 2018

    work page 1994

  2. [2]

    Mart ´ın Abadi, Ashish Agarwal, Paul Barham, Eugene Brevdo, Zhifeng Chen, Craig Citro, Greg S. Corrado, Andy Davis, Jeffrey Dean, Matthieu Devin, Sanjay Ghemawat, Ian Goodfellow, Andrew Harp, Geoffrey Irving, Michael Isard, Yangqing Jia, Rafal Jozefowicz, Lukasz Kaiser, Manjunath Kudlur, Josh Levenberg, Dandelion Man ´e, Rajat Monga, Sherry Moore, Derek M...

    work page 2015

  3. [3]

    Atapour-Abarghouei and T.P

    A. Atapour-Abarghouei and T.P. Breckon. Real-time monoc- ular depth estimation using synthetic data with domain adap- tation. In IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pages 1–8, 2018

    work page 2018

  4. [4]

    R. O. Castle, D. J. Gawley, G. Klein, and D. W. Murray. Towards simultaneous recognition, localization and mapping for hand-held and wearable cameras. In ”International Con- ference on Robotics and Automation (ICRA)” , pages 4102– 4107, 2007

    work page 2007

  5. [5]

    Long-term recurrent convolutional net- works for visual recognition and description

    Jeff Donahue, Lisa Anne Hendricks, Sergio Guadarrama, Marcus Rohrbach, Subhashini Venugopalan, Kate Saenko, and Trevor Darrell. Long-term recurrent convolutional net- works for visual recognition and description. In Proceed- ings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pages 2625–2634, 2015

    work page 2015

  6. [6]

    Dosovitskiy, P

    A. Dosovitskiy, P. Fischer, E. Ilg, P. H¨ausser, C. Hazırbas ¸, V . Golkov, P. v.d. Smagt, D. Cremers, and T. Brox. Flownet: Learning optical flow with convolutional networks. In Pro- ceedings of the IEEE International Conference on Computer Vision (ICCV), pages 2758–2766, 2015

    work page 2015

  7. [7]

    CARLA: An open urban driving simulator

    Alexey Dosovitskiy, German Ros, Felipe Codevilla, Antonio Lopez, and Vladlen Koltun. CARLA: An open urban driving simulator. In Proceedings of the 1st Annual Conference on Robot Learning, pages 1–16, 2017

    2017

  8. [8]

    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. In Advances in Neural Information Processing Sys- tems, pages 2366–2374. 2014

    work page 2014

Show all 36 references
  1. [9]

    Frost, Olaf K ¨ahler, and David W

    Duncan P. Frost, Olaf K ¨ahler, and David W. Murray. Object- aware bundle adjustment for correcting monocular scale drift. In ”International Conference on Robotics and Automa- tion (ICRA)”, pages 4770–4776, 2016

    2016

  2. [10]

    Frost, David W

    Duncan P. Frost, David W. Murray, and Victor Adrian Prisacariu. Using learning of speed to stabilize scale in monocular localization and mapping. In International Con- ference on 3D Vision (3DV), pages 527–536, 2017

    2017

  3. [11]

    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 (CVPR), pages 2002–2011, 2018

    2002

  4. [12]

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

    Andreas Geiger. Are we ready for autonomous driving? the kitti vision benchmark suite. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pages 3354–3361, 2012

    2012

  5. [13]

    3d traffic scene understanding from movable platforms

    Andreas Geiger, Martin Lauer, Christian Wojek, Christoph Stiller, and Raquel Urtasun. 3d traffic scene understanding from movable platforms. IEEE transactions on pattern anal- ysis and machine intelligence, 36(5):1012–1025, 2013

    2013

  6. [14]

    Stere- oscan: Dense 3d reconstruction in real-time

    Andreas Geiger, Julius Ziegler, and Christoph Stiller. Stere- oscan: Dense 3d reconstruction in real-time. In IEEE Intel- ligent V ehicles Symposium (IV), pages 963–968, 2011

    2011

  7. [15]

    Robust scale estimation for monocular visual odometry using struc- ture from motion and vanishing points

    Johannes Gr ¨ater, Tobias Schwarze, and Martin Lauer. Robust scale estimation for monocular visual odometry using struc- ture from motion and vanishing points. In IEEE Intelligent V ehicles Symposium (IV), pages 475–480, 2015

    2015

  8. [16]

    Long short-term memory

    Sepp Hochreiter and J ¨urgen Schmidhuber. Long short-term memory. Neural Computation, 9(8):1735–1780, 1997

    1997

  9. [17]

    Multimodal unsupervised image-to-image translation

    Xun Huang, Ming-Yu Liu, Serge Belongie, and Jan Kautz. Multimodal unsupervised image-to-image translation. InEu- ropean Conference on Computer Vision (ECCV), pages 172– 189, 2018

    2018

  10. [18]

    Kinectfusion: Real-time 3d recon- struction and interaction using a moving depth camera

    Shahram Izadi, David Kim, Otmar Hilliges, David Molyneaux, Richard Newcombe, Pushmeet Kohli, Jamie Shotton, Steve Hodges, Dustin Freeman, Andrew Davison, and Andrew Fitzgibbon. Kinectfusion: Real-time 3d recon- struction and interaction using a moving depth camera. In Proceedi...

    2011

  11. [19]

    C. Kerl, J. Sturm, and D. Cremers. Robust odometry esti- mation for rgb-d cameras. In ”International Conference on Robotics and Automation (ICRA)” , pages 3748–3754, 2013

    2013

  12. [20]

    Adam: A method for stochastic optimization

    Diederik P Kingma and Jimmy Ba. Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980 , 2014

    2014 arXiv

  13. [21]

    Vi- sual odometry based on stereo image sequences with ransac- based outlier rejection scheme

    Bernd Kitt, Andreas Geiger, and Henning Lategahn. Vi- sual odometry based on stereo image sequences with ransac- based outlier rejection scheme. In IEEE Intelligent V ehicles Symposium (IV), pages 486–492, 2010

    2010

  14. [22]

    Rela- tive geometry-aware siamese neural network for 6dof camera relocalization

    Qing Li, Jiasong Zhu, Rui Cao, Ke Sun, Jonathan M Garibaldi, Qingquan Li, Bozhi Liu, and Guoping Qiu. Rela- tive geometry-aware siamese neural network for 6dof camera relocalization. arXiv preprint arXiv:1901.01049, 2019

    1901 arXiv

  15. [23]

    What makes good synthetic training data for learning dispar- ity and optical flow estimation? International Journal of Computer Vision, 126(9):942–960, 2018

    Nikolaus Mayer, Eddy Ilg, Philipp Fischer, Caner Hazir- bas, Daniel Cremers, Alexey Dosovitskiy, and Thomas Brox. What makes good synthetic training data for learning dispar- ity and optical flow estimation? International Journal of Computer Vision, 126(9):942–960, 2018

    2018

  16. [24]

    deepdrive/deepdrive: 2.0, 2018

    Craig Quiter and Maik Ernst. deepdrive/deepdrive: 2.0, 2018

    2018

  17. [25]

    Three-dimensional absolute orientation of stereo models using digital elevation models

    Dan Rosenholm and Kenner! Torlegard. Three-dimensional absolute orientation of stereo models using digital elevation models. Photogrammetric Engineering and Remote Sensing, 54(10):4102–4107, 1988

    1988

  18. [26]

    Absolute scale in structure from mo- tion from a single vehicle mounted camera by exploiting nonholonomic constraints

    Davide Scaramuzza, Friedrich Fraundorfer, Marc Pollefeys, and Roland Siegwart. Absolute scale in structure from mo- tion from a single vehicle mounted camera by exploiting nonholonomic constraints. In Proceedings of the IEEE In- ternational Conference on Computer Vision (ICCV)...

    2009

  19. [27]

    Absolute scale in structure from mo- tion from a single vehicle mounted camera by exploiting nonholonomic constraints

    Davide Scaramuzza, Friedrich Fraundorfer, Marc Pollefeys, and Roland Siegwart. Absolute scale in structure from mo- tion from a single vehicle mounted camera by exploiting nonholonomic constraints. In Proceedings of the Interna- tional Conference on Computer Vision (ICCV), pag...

    2009

  20. [28]

    Airsim: High-fidelity visual and physical simula- tion for autonomous vehicles

    Shital Shah, Debadeepta Dey, Chris Lovett, and Ashish Kapoor. Airsim: High-fidelity visual and physical simula- tion for autonomous vehicles. In Field and Service Robotics, pages 621–635, 2017

    2017

  21. [29]

    Robust scale esti- mation in real-time monocular SFM for autonomous driving

    Shiyu Song and Manmohan Chandraker. Robust scale esti- mation in real-time monocular SFM for autonomous driving. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pages 1566–1573, 2014

    2014

  22. [30]

    Bayesian scale estima- tion for monocular SLAM based on generic object detection for correcting scale drift

    Edgar Sucar and Jean-Bernard Hayet. Bayesian scale estima- tion for monocular SLAM based on generic object detection for correcting scale drift. In ”International Conference on Robotics and Automation (ICRA)” , pages 1–7, 2018

    2018

  23. [31]

    Tateno, F

    K. Tateno, F. Tombari, I. Laina, and N. Navab. Cnn-slam: Real-time dense monocular slam with learned depth predic- tion. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR) , pages 6243–6252, 2017

    2017

  24. [32]

    S. Wang, R. Clark, H. Wen, and N. Trigoni. Deepvo: To- wards end-to-end visual odometry with deep recurrent con- volutional neural networks. pages 2043–2050, 2017

    2017

  25. [33]

    N. Yang, R. Wang, J. Stueckler, and D. Cremers. Deep vir- tual stereo odometry: Leveraging deep depth prediction for monocular direct sparse odometry. In Proceedings of the IEEE European Conference on Computer Vision (ECCV) , pages 817–833, 2018

    2018

  26. [34]

    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 Proceedings of the IEEE Euro- pean Conference on Computer Vision (ECCV) , pages 767– 783, 2018

    2018

  27. [35]

    Ground plane based absolute scale estimation for monocular visual odom- etry

    Dingfu Zhou, Yuchao Dai, and Hongdong Li. Ground plane based absolute scale estimation for monocular visual odom- etry. arXiv preprint arXiv:1903.00912, 2019

    1903 arXiv

  28. [36]

    Unpaired image-to-image translation using cycle- consistent adversarial networks

    Jun-Yan Zhu, Taesung Park, Phillip Isola, and Alexei A Efros. Unpaired image-to-image translation using cycle- consistent adversarial networks. In IEEE International Conference on Computer Vision (ICCV) , pages 2223–2232, 2017

    2017

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.