Pith. sign in

REVIEW 4 major objections 5 minor 24 references

Self-Supervised Monocular Visual Drone Model Identification through Improved Occlusion Handling

T0 review · 4 major / 5 minor · reviewed 2026-08-16 · deepseek-v4-flash

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

Pith's one-line read A neural drone model trained without ground truth beats its vision teacher at high speed and improves filter-based odometry on aggressive tracks.

desk verdict The self-supervised drone model is a genuine new capability, but the student-beats-teacher claim at high speed rests on relative error plotted against speed, with no absolute-error check, so the paper needs revision before the headline is credible. read the letter →

arxiv 2504.21695 v1 pith:23HO7CAW submitted 2025-04-30 cs.RO cs.AI

Stavrow A. Bahnam , Christophe De Wagter , Guido C.H.E. de Croon This is my paper · ORCID
classification cs.ROcs.AI
keywords self-supervisedlearningmonocularvisualodometrydronemodelidentificationocclusionhandlingvisual-inertialracingteacher-studenttrainingego-motionestimation
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 tries to remove the need for external motion capture when learning a drone's flight dynamics. It proposes a two-stage self-supervised scheme: first train a monocular pose network (PoseNet) on onboard video alone, then use that network's velocity estimates as a teacher for a small neural drone model that takes inertial measurement unit (IMU) readings and motor RPM feedback as inputs. The paper introduces an occlusion-handling loss that masks invalid pixels and takes the minimum of two reprojection errors, which makes the teacher more accurate near racing gates. The central result is that the learned drone model becomes more accurate than its vision teacher at high speeds, and that blending it into a filter-based visual-inertial odometry system improves position estimates on aggressive 3D racing trajectories. If the approach holds, drone ego-motion estimation could scale to any drone in any GPS-denied environment without ground-truth infrastructure.

What carries the argument

Three mechanisms carry the argument. First, the teacher-student setup: a self-supervised PoseNet (a monocular relative-pose network trained by image reconstruction) provides unscaled velocity estimates, and a single learnable scale parameter per training sequence recovers metric scale by forcing the integrated drone model velocity to match. Second, the specific-force parameterization in Equation (10): the drone model, a three-layer fully connected network, outputs two drag coefficients $d_x,d_y$ in $[0,2]$ and an accelerometer residual $\varepsilon_{az}$ in $[-5,5]$; body velocity is obtained by integrating gravity-compensated specific forces, where lateral acceleration is $-d_x V_{bx}$ and $-d_y V_{by}$. The occlusion-handling loss is the third enabling piece: for both photometric and depth consistency, the loss takes the minimum of two reprojection errors but only over pixels marked valid by a mask, preventing erroneous gradients from pixels that disappear behind gates or leave the image. This makes the PoseNet teacher sharper and more scale-consistent, which in turn makes the student's supervision signal reliable.

What would settle it

Run the same self-supervised pipeline on flights with strong wind or with rapid yaw rotations at constant linear speed, compare the open-loop integrated velocity of the learned drone model against motion-capture ground truth, and check whether the relative velocity error at 15–20 m/s stays below the PoseNet teacher's error; if instead the model drifts or the drag coefficients change systematically with yaw direction, Equation (10)'s assumed dynamics are incomplete and the central claim fails.

Watch

Extended reading notes

Core claim

On the paper's own terms, the discovery is that a quadrotor's velocity can be learned from onboard sensors alone, without any ground-truth labels, through a teacher-student scheme in which a self-supervised monocular pose network supervises a neural dynamics model. The improved occlusion handling combines a valid-pixel mask with minimum reprojection error in both the photometric loss and the depth consistency loss; the two-frame variant, which reprojects two frames onto each other with a transformation and its inverse, reduces the PoseNet's absolute position RMSE by an average of 15% relative to the benchmark. The student drone model predicts specific forces: lateral accelerations equal to learned drag coefficients times body velocity, and a vertical acceleration equal to the measured accelerometer z-value minus a learned residual. Integrated over time, this model's relative velocity error decreases as speed increases, whereas the teacher's error increases, and the model is robust to the motion blur that degrades vision during aggressive rotation. When 30% of the drone model's acceleration prediction is blended with 70% of the IMU measurement inside a filter-based visual-inertial odometry system (ROVIO), absolute position RMSE improves on aggressive 3D racing trajectories, especially when the image processing rate is reduced.

Load-bearing premise

The scheme rests on the assumption that a quadrotor's sideways acceleration is exactly a coefficient times its sideways velocity, that thrust points exactly along the body z-axis, and that the measured accelerometer z-value minus a bounded learned residual captures all vertical aerodynamic effects; if real-world couplings such as uneven rotor airspeed during rotation, blade flapping, wake turbulence, or wind violate this form, the integrated velocity drifts and the claimed high-speed advantage over the vision teacher will not generalize.

Editorial extensions

If this is right

  • The 2F occlusion-handling loss reduces the self-supervised PoseNet's absolute position RMSE by an average of 15% relative to the benchmark network and by 25% relative to the 3F variant.
  • A neural drone model can be trained entirely from onboard monocular video, IMU data, and motor RPM feedback, removing motion capture from the training loop.
  • The learned drone model's relative velocity error decreases with flight speed while PoseNet's increases, so the model is the more reliable velocity source during fast, rotation-heavy flight.
  • Blending 30% drone model acceleration with 70% IMU acceleration inside filter-based visual-inertial odometry lowers absolute position RMSE on aggressive 3D racing trajectories, with the largest gains at reduced image processing rates.
  • Sharper depth maps near gates follow from the occlusion-handling loss, which makes the teacher's translation estimates more scale-consistent and thereby improves student training.

Reading between the lines

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

  • The same teacher-student recipe could transfer to other airframes or even ground vehicles, since the scale-recovery mechanism only requires an IMU attitude estimate and known gravity, not a vision-specific scale prior.
  • The occlusion loss is not tied to drone racing: any monocular depth or ego-motion network facing large disocclusions, such as urban scenes with poles and signs, could adopt the valid-mask plus minimum-of-two-reprojection scheme.
  • If the learned model stays accurate open-loop, a natural next step is to run it during complete visual dropout or at very low camera rates, effectively shifting from visual-inertial odometry toward model-based inertial odometry with occasional visual resets.
  • A concrete test the paper does not perform is cross-drone transfer: training the student on one quadrotor and evaluating on another would reveal whether the drag parameterization captures general quadrotor physics or only the training airframe's specifics.
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

4 major / 5 minor

Summary. The paper proposes a fully self-supervised pipeline for drone ego-motion estimation. First, a monocular PoseNet/DepthNet is trained with modified photometric and depth-consistency losses that combine minimum-reprojection and valid-pixel masking to handle occlusions and out-of-view pixels. The PoseNet then serves as a teacher for a small neural-network drone model that predicts specific forces from IMU and motor-RPM data, with a learnable scale parameter resolving the monocular scale ambiguity. The learned drone model is finally fused with ROVIO by combining its predicted acceleration with the IMU accelerometer. The paper reports that the proposed occlusion handling reduces absolute position RMSE by 15% on the TII Drone Racing Dataset, that the student drone model outperforms its PoseNet teacher in relative velocity error at high speeds, and that fusing the drone model into ROVIO improves absolute position RMSE on aggressive 3D racing trajectories.

Significance. If the claims hold, the contribution is valuable: it removes the motion-capture ground-truth requirement for learning a drone dynamics model, which is a practical limitation of prior work such as Cioffi et al. and Swift. The paper also uses a public high-speed drone dataset, evaluates against motion-capture ground truth, and integrates the learned model into a standard VIO filter, which is a concrete and useful demonstration. The strengths are the self-supervised formulation, the public benchmark, and the clear integration experiment. However, the headline comparisons are currently not fully supported by the reported numbers and plots: the occlusion-handling improvement is overstated relative to Table I, and the student-vs-teacher velocity comparison relies on a relative-error metric that can create an artifactual trend. These issues are fixable with additional analysis, so the underlying approach remains credible.

major comments (4)
  1. [III-B, Table I] The text states that the 2F method 'achieves a lower RMSE than the 3F approach for nearly all trajectories, with an average reduction of 25%.' This is not supported by Table I. The average RMSE for 3F is 2.51 m and for 2F is 2.00 m, which is a reduction of about 20%, not 25%. Moreover, 2F is worse than 3F on 7 of the 15 reported rows (12P, 01-04A, 05A, 07-10A, 12A, 13-16A, 18A), so 'nearly all trajectories' is inaccurate. Please correct the numbers and the qualitative claim, and report per-trajectory statistics or a paired test to support the comparison.
  2. [III-C, Figure 6] The central claim that the student drone model becomes more accurate than its teacher at higher speeds rests on Figure 6, which plots relative velocity error. This metric is problematic here: the student is evaluated by open-loop integration initialized at zero velocity, so low-speed samples near the start naturally have large relative errors, while high-speed samples later in the trajectory have smaller relative errors even if the absolute error is flat or growing. A model with a roughly constant absolute bias automatically shows decreasing relative error with speed. The paper does not report absolute velocity error in m/s, per-bin sample counts, or confidence intervals for Figure 6, so the plotted crossover may be an artifact of normalization. Please add an absolute-error comparison, with speed-bin statistics, for both teacher and student; if the student's absolute error at high speeds is not lower than the teacher's, the headline claim and the motivation for the ROVIO integration are not directly supported.
  3. [II-C, Eq. (10)] The learned drone model assumes a restrictive specific-force parameterization: lateral specific forces are exactly -dx Vbx and -dy Vby with dx, dy in [0,2], the z-axis specific force is the measured accelerometer value minus a residual bounded in [-5,5], and thrust is assumed to lie exactly on the body z-axis. Coupled effects such as yaw-rate-dependent rotor drag, blade flapping, wake turbulence, and wind are excluded. The experiments are all on the TII indoor dataset, so the model may only fit this distribution. Since the abstract and conclusion claim applicability to 'any drone in any environment,' this assumption is load-bearing for generalization. Please either provide evidence that the parameterization captures the relevant dynamics outside the training distribution, or explicitly narrow the generality claims and discuss the expected failure modes.
  4. [III-C, Figure 7] The ROVIO integration result uses a fixed fusion weight of 30% neural model and 70% IMU acceleration, but no sensitivity analysis or selection criterion is reported. The improvement is shown mainly for the autonomous 3D race track; on the 2D tracks the hybrid method does not appear consistently better. To support the claim that the learned drone model 'significantly improves state estimation,' please report per-trajectory RMSE for the test set, the effect of varying the fusion weight, and a statistical comparison rather than boxplot means alone.
minor comments (5)
  1. [II-B, Figures 2-3] The notation for the schemes is confusing: Figure 2's caption calls the min-per-valid-pixel combination of two reprojections the '3F scheme,' while Section II-B later defines '3F' and '2F' as two alternative schemes and the experiments compare them as such. Please unify the terminology so that the proposed losses, the two schemes, and the benchmark are unambiguously defined.
  2. [III-B, Table I] The abstract says the occlusion-handling method reduces RMSE by an average of 15%, which matches the 2F-vs-benchmark comparison (2.35 to 2.00), but the same table shows the 3F method is worse than the benchmark on average. Please clarify in the text and abstract which method is 'the proposed method' and which comparison the 15% refers to.
  3. [III-C, Figure 6] The teacher velocity is computed by dividing PoseNet translation by the frame interval, while the student integrates accelerations; these are different estimation tasks. The paper also says a Butterworth filter is applied to PoseNet estimates before training, but it is unclear whether Figure 6 evaluates the filtered or unfiltered teacher. Please state explicitly which version is plotted and whether the conclusion survives comparing the student against the same smoothed teacher.
  4. [II-C, Eq. (8)] Equation (8) uses a camera-to-body transformation R_{c->b}, but the text says camera-IMU extrinsics are required. Please clarify whether this is a rotation-only transform and how the translation between camera and IMU is handled, since that affects the body-frame velocity computation.
  5. [Figure 5] The x-axis label 'Timestamp [s] +1.69175617e9' appears to contain an unformatted epoch offset. Please remove or format this artifact.

Circularity Check

0 steps flagged · score 0.0 of 10

No circular dependency found: the student-drone-model comparison is evaluated against motion-capture ground truth, and the scale-ambiguity calibration is anchored to IMU/gravity dynamics rather than to the claims being tested.

full rationale

The paper's derivation chain is not circular. The self-supervised PoseNet teacher is trained with photometric, depth-consistency, and smoothness losses and evaluated by absolute position RMSE against ground truth after SIM3 alignment (Section III-B, Table I), which is an independent external benchmark. The drone-model student is trained using PoseNet velocity labels with per-sequence scale parameters s (Eq. 8), but the scale is calibrated jointly with the dynamics model: the student integrates specific forces plus the known gravity vector (Eq. 10), so the metric scale is anchored to the IMU accelerometer and 9.81 m/s^2 rather than to the claimed outcome. The student's velocity accuracy in Figure 6 is obtained by open-loop integration from zero initial velocity and compared to motion-capture ground truth, and the ROVIO experiments in Figure 7 are likewise evaluated against ground-truth trajectories; neither comparison reduces to the training labels by construction. The cited works by the same group ([3], [5], [7], [21]) are background or domain-knowledge references (e.g., rotor-drag linearity, prior AlphaPilot model-based odometry) and are not invoked to force the paper's central result. The relative-error metric in Figure 6 is a possible evaluation weakness, but that is a correctness risk, not a circular dependence.

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

The central method rests on a handful of modeling assumptions (indoor, wind-free, thrust aligned with body z-axis, known extrinsics) and several hand-picked hyperparameters (loss weights, output bounds, 30/70 fusion weight). The most consequential free parameter is the per-sequence scale s that aligns the scale-ambiguous teacher, plus the fusion weight that drives the headline ROVIO result.

free parameters (4)
  • PoseNet scale parameter s (per sequence) = learned per training sequence, values not reported
    Equation 8 applies a single learnable scale per sequence to convert scale-ambiguous PoseNet translations into metric body velocities; it is fitted jointly with the drone model and is essential for the teacher-student training.
  • Loss weights λ1, λ2 = λ1=0.15, λ2=0.001
    Set by hand in Equation 7 for depth consistency and smoothness losses; affects the PoseNet training quality.
  • ROVIO fusion weight for neural drone model acceleration = 0.30 (30% neural, 70% IMU)
    Chosen by hand in Section III-C; the reported improvement in Figure 7 depends on this weighting, and no sensitivity analysis is provided.
  • Output bounds for drag terms and accelerometer residual = dx, dy ∈ [0,2]; ε_accz ∈ [-5,5]
    Hand-chosen sigmoid scaling limits in the drone model output layer, constraining the learned dynamics.
assumptions (5)
  • domain assumption External wind disturbances are negligible (indoor environment).
    Stated in Section II-C before Equation 9; if false, drag estimates would absorb unmodeled wind and the model would not transfer outdoors.
  • domain assumption Thrust is aligned exactly with the body z-axis.
    Section II-C assumes a quadrotor with thrust along body z; real drones have thrust vectoring and blade flapping that violate this, particularly in aggressive turns.
  • domain assumption IMU is located in the body frame and camera-IMU extrinsics are known.
    Required to transform camera-frame PoseNet velocities to body frame in Equation 8; in practice extrinsics need calibration.
  • domain assumption PoseNet scale is consistent within each trajectory and can be fixed by one scalar per sequence.
    Section II-C: 'PoseNet's velocity estimates should be proportional to the true scale by a single unknown factor'; the per-sequence s fit is only valid if scale is internally consistent.
  • domain assumption BetaFlight-style EKF attitude estimate provides an accurate gravity direction.
    Gravity vector g is computed from the attitude EKF; errors directly bias the integrated velocity in Equation 10.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Self-Supervised Monocular Visual Drone Model Identification through Improved Occlusion Handling." pith.science (2026). https://pith.science/paper/23HO7CAW

@misc{pith2026250421695,
  author       = {Pith},
  title        = {Pith review of: Self-Supervised Monocular Visual Drone Model Identification through Improved Occlusion Handling},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/23HO7CAW}},
  note         = {Machine review of arXiv:2504.21695}
}
read the original abstract

Ego-motion estimation is vital for drones when flying in GPS-denied environments. Vision-based methods struggle when flight speed increases and close-by objects lead to difficult visual conditions with considerable motion blur and large occlusions. To tackle this, vision is typically complemented by state estimation filters that combine a drone model with inertial measurements. However, these drone models are currently learned in a supervised manner with ground-truth data from external motion capture systems, limiting scalability to different environments and drones. In this work, we propose a self-supervised learning scheme to train a neural-network-based drone model using only onboard monocular video and flight controller data (IMU and motor feedback). We achieve this by first training a self-supervised relative pose estimation model, which then serves as a teacher for the drone model. To allow this to work at high speed close to obstacles, we propose an improved occlusion handling method for training self-supervised pose estimation models. Due to this method, the root mean squared error of resulting odometry estimates is reduced by an average of 15%. Moreover, the student neural drone model can be successfully obtained from the onboard data. It even becomes more accurate at higher speeds compared to its teacher, the self-supervised vision-based model. We demonstrate the value of the neural drone model by integrating it into a traditional filter-based VIO system (ROVIO), resulting in superior odometry accuracy on aggressive 3D racing trajectories near obstacles. Self-supervised learning of ego-motion estimation represents a significant step toward bridging the gap between flying in controlled, expensive lab environments and real-world drone applications. The fusion of vision and drone models will enable higher-speed flight and improve state estimation, on any drone in any environment.

Figures

Figures reproduced from arXiv: 2504.21695 by the authors.

Figure 1
Figure 1. We introduce a self-supervised learning approach that [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Validity image reprojection: Given three images at timestep t − 5, t and t + 5, two reconstructions of image It can be computed. By estimating a depth map at time t and two translation and rotation estimations, Tt−5→t and Tt+5→t. In the occluded parts, this reconstruction is undefined and creates artefacts. For instance, in the reconstructed image that uses image t − 5 as input (top row), the pixels right and below … view at source ↗
Figure 3
Figure 3. Validity depth reprojection: When approaching a gate, the reprojected depth map perceives similar artefacts as reprojected images due to occlusion. This causes an error in the depth consistency loss. In [14] only one projection was considered and used the depth consistency loss based on the top Error Map. In our 2F method we reprojected Dt as well (bottom row) using the inverse of the estimated relative pose (T), fo… view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Single learnable parameter (in T) is used to transform the scaleless PoseNet velocity to a scaled body velocity. This [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]
Figure 5
Figure 5. Figure 5: Our loss function improves both depth and velocity [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]
Figure 7
Figure 7. Figure 7: Absolute RMSE (after 6DOF alginment) vs. Skipping Frames Using Traditional ROVIO and a Neural Drone Model [PITH_FULL_IMAGE:figures/full_fig_p008_7.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

24 extracted references · 19 canonical work pages

  1. [1]

    Visual-inertial navigation: A concise review,

    G. Huang, “Visual-inertial navigation: A concise review,” in 2019 International Conference on Robotics and Automation (ICRA) , 2019, pp. 9572–9582

    work page 2019

  2. [2]

    Iterated extended kalman filter based visual-inertial odometry using direct pho- tometric feedback,

    M. Bloesch, M. Burri, S. Omari, M. Hutter, and R. Siegwart, “Iterated extended kalman filter based visual-inertial odometry using direct pho- tometric feedback,” The International Journal of Robotics Research , vol. 36, pp. 1053–1072, 09 2017

    work page 2017

  3. [3]

    Cnn-based ego-motion estimation for fast mav maneuvers,

    Y . Xu and G. C. de Croon, “Cnn-based ego-motion estimation for fast mav maneuvers,” in 2021 IEEE International Conference on Robotics and Automation (ICRA) . IEEE, 2021, pp. 7606–7612

    work page 2021

  4. [4]

    A self-supervised, differentiable kalman filter for uncertainty-aware visual-inertial odometry,

    B. Wagstaff, E. Wise, and J. Kelly, “A self-supervised, differentiable kalman filter for uncertainty-aware visual-inertial odometry,” in 2022 IEEE/ASME International Conference on Advanced Intelligent Mecha- tronics (AIM), 2022, pp. 1388–1395

    work page 2022

  5. [5]

    Cuahn-vio: Content-and-uncertainty- aware homography network for visual-inertial odometry,

    Y . Xu and G. C. de Croon, “Cuahn-vio: Content-and-uncertainty- aware homography network for visual-inertial odometry,”Robotics and Autonomous Systems , vol. 185, p. 104866, 2025

    work page 2025

  6. [6]

    Alphapilot: autonomous drone racing,

    P. Foehn, D. Brescianini, E. Kaufmann, T. Cieslewski, M. Gehrig, M. Muglikar, and D. Scaramuzza, “Alphapilot: autonomous drone racing,” Autonomous Robots , vol. 46, no. 1, pp. 307–320, 2022. [Online]. Available: https://doi.org/10.1007/s10514-021-10011-y

    doi:10.1007/s10514-021-10011-y 2022

  7. [7]

    The sensing, state-estimation, and control behind the winning entry to the 2019 artificial intelligence robotic racing competition,

    C. De Wagter, F. Paredes-Vall ´es, N. Sheth, and G. Croon, “The sensing, state-estimation, and control behind the winning entry to the 2019 artificial intelligence robotic racing competition,” Field Robotics, vol. 2, pp. 1263–1290, 03 2022

    work page 2019

  8. [8]

    Learned inertial odometry for autonomous drone racing,

    G. Cioffi, L. Bauersfeld, E. Kaufmann, and D. Scaramuzza, “Learned inertial odometry for autonomous drone racing,” IEEE Robotics and Automation Letters, vol. PP, pp. 1–8, 05 2023

    work page 2023

Show all 24 references
  1. [9]

    Champion-level drone racing using deep rein- forcement learning,

    E. Kaufmann, L. Bauersfeld, A. Loquercio, M. Mueller, V . Koltun, and D. Scaramuzza, “Champion-level drone racing using deep rein- forcement learning,” Nature, vol. 620, pp. 982–987, 08 2023

    2023

  2. [10]

    Intel realsense t265 series product family,

    “Intel realsense t265 series product family,” https://www.intelrealsense. com/wp-content/uploads/2019/09/Intel RealSense Tracking Camera Datasheet Rev004 release.pdf, 2019, accessed: 2024-09-15

    2019

  3. [11]

    Demonstrating agile flight from pixels without state estimation,

    I. Geles, L. Bauersfeld, A. Romero, J. Xing, and D. Scaramuzza, “Demonstrating agile flight from pixels without state estimation,” arXiv preprint arXiv:2406.12505 , 2024

    2024 arXiv

  4. [12]

    Digging into self- supervised monocular depth estimation,

    C. Godard, O. Aodha, M. Firman, and G. Brostow, “Digging into self- supervised monocular depth estimation,” in 2019 IEEE/CVF Interna- tional Conference on Computer Vision (ICCV) , 11 2019, pp. 3827– 3837

    2019

  5. [13]

    Unsupervised learning of depth and ego-motion from monocular video using 3d geometric constraints,

    R. Mahjourian, M. Wicke, and A. Angelova, “Unsupervised learning of depth and ego-motion from monocular video using 3d geometric constraints,” in 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2018, pp. 5667–5675

    2018

  6. [14]

    Unsupervised scale-consistent depth and ego-motion learning from monocular video,

    J. Bian, Z. Li, N. Wang, H. Zhan, C. Shen, M.-M. Cheng, and I. Reid, “Unsupervised scale-consistent depth and ego-motion learning from monocular video,” in Advances in Neural Information Processing Systems, H. Wallach, H. Larochelle, A. Beygelzimer, F. d'Alch ´e-Buc, E. Fox, a...

    2019

  7. [15]

    U-net: Convolutional networks for biomedical image segmentation,

    O. Ronneberger, P. Fischer, and T. Brox, “U-net: Convolutional networks for biomedical image segmentation,” in Medical Image Computing and Computer-Assisted Intervention – MICCAI 2015 , vol

    2015

  8. [16]

    Deep residual learning for image recognition,

    K. He, X. Zhang, S. Ren, and J. Sun, “Deep residual learning for image recognition,” in 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) , 06 2016, pp. 770–778

    2016

  9. [17]

    Image quality assessment: from error visibility to structural similarity,

    Z. Wang, A. Bovik, H. Sheikh, and E. Simoncelli, “Image quality assessment: from error visibility to structural similarity,” IEEE Trans- actions on Image Processing , vol. 13, no. 4, pp. 600–612, 2004

    2004

  10. [18]

    Unsupervised monocular depth estimation with left-right consistency,

    C. Godard, O. Aodha, and G. Brostow, “Unsupervised monocular depth estimation with left-right consistency,” in2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) , 07 2017

    2017

  11. [19]

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

    A. Geiger, P. Lenz, and R. Urtasun, “Are we ready for autonomous driving? the kitti vision benchmark suite,” in Conference on Computer Vision and Pattern Recognition (CVPR) , 2012

    2012

  12. [20]

    The betaflight open source flight controller firmware project. betafligh,

    “The betaflight open source flight controller firmware project. betafligh,” 2024, computer software. [Online]. Available: https: //github.com/betaflight/betaflight

    2024

  13. [21]

    Quadrotor thrust vectoring control with time and jerk optimal trajectory planning in constant wind fields,

    J. P. Silva, C. De Wagter, and G. de Croon, “Quadrotor thrust vectoring control with time and jerk optimal trajectory planning in constant wind fields,” Unmanned Systems , vol. 6, no. 01, pp. 15–37, 2018

    2018

  14. [22]

    Race against the machine: A fully-annotated, open-design dataset of autonomous and piloted high-speed flight,

    M. Bosello, D. Aguiari, Y . Keuter, E. Pallotta, S. Kiade, G. Caminati, F. Pinzarrone, J. Halepota, J. Panerati, and G. Pau, “Race against the machine: A fully-annotated, open-design dataset of autonomous and piloted high-speed flight,” IEEE Robotics and Automation Letters , v...

    2024

  15. [23]

    Least-squares estimation of transformation parameters between two point patterns,

    S. Umeyama, “Least-squares estimation of transformation parameters between two point patterns,” IEEE Transactions on Pattern Analysis and Machine Intelligence , vol. 13, no. 4, pp. 376–380, 1991

    1991

  16. [9351]

    Springer International Publishing, 10 2015, pp. 234–241

    2015

Pith tools

Reviewed August 16, 2026 · model on record in the stance chip above.