Pith. sign in

REVIEW 4 major objections 6 minor 1 cited by

TopView: Vectorising road users in a bird's eye view from uncalibrated street-level imagery with deep learning

T0 review · 4 major / 6 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 This paper claims that a vector bird's-eye map of road users can be generated from a single uncalibrated street-level image or video by learning the scene's vanishing point, without knowing camera intrinsics or extrinsics.

desk verdict A genuinely new pipeline for vector BEV maps from uncalibrated imagery, but the metric geolocation claim doesn't survive the geometry and the quantitative comparison is unexplained. read the letter →

arxiv 2412.16229 v1 pith:2DMOXMWB submitted 2024-12-18 cs.CV

Mohamed R Ibrahim This is my paper · ORCID
classification cs.CV
keywords bird'seyeviewhomographydeeplearningurbanscenesvanishingpointobjectdetectionuncalibratedcameratrajectorymapping
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 claims that a vector bird's-eye map of road users can be generated from a single uncalibrated street-level image or video by learning the scene's vanishing point, without knowing the camera's intrinsic or extrinsic parameters. The core idea is to regress the vanishing point with a deep network and then use that point to construct a homography that projects detected road users onto a top-down plane. The paper also derives pseudo-3D bounding boxes and orientation labels deterministically from 2D boxes, the vanishing point, and trajectory lines. If correct, any CCTV feed or internet image becomes convertible into a metric, map-aligned representation of traffic and pedestrians, enabling distance measurement, conflict detection, and city-scale occupancy analytics without per-camera calibration.

What carries the argument

The central object is the learned vanishing point of the scene, combined with a homography built from it. The vanishing point is regressed by a MobileNet-based network trained on 172,576 images from six datasets. The homography module draws a horizontal line near the bottom of the image, subdivides it uniformly, and connects those subdivisions to the vanishing point; the intersections of these radial lines with upper and lower horizontal bounds define a four-point quadrilateral that is mapped to a BEV rectangle. Algorithm 1 then converts a 2D box plus trajectory line into a 3D box by comparing the trajectory line's intersection with the box's top edge against the midpoint and the vanishing point's offset from image center, yielding orientation tags such as 'turning left', 'moving straight', or 'side view'.

What would settle it

Take a calibrated camera on a flat, level ground plane with known intrinsics and extrinsics, photograph or render a scene, run TopView, and compare the BEV coordinates of objects against the ground-truth metric positions from the calibration. Systematically larger errors for objects farther from the bottom-center of the frame, or for scenes with a visible camera roll such as a tilted horizon, would falsify the claim that a vanishing point plus two scalars is sufficient for metric BEV reconstruction.

Watch

Extended reading notes

Core claim

The central claim is that an orthogonal (bird's-eye) projection of road users can be estimated purely from the geometry of a single uncalibrated image by learning its vanishing point. TopView regresses the vanishing point and horizon from the image, detects road users with a YOLO detector, and then constructs a quadrilateral in the image plane—formed by radial lines from the vanishing point to a horizontal line at the bottom of the frame—that is mapped by homography to a rectangle in the BEV vector space. The same vanishing point, together with an object's trajectory line, constrains a 3D bounding box inside each 2D box via a deterministic geometric algorithm. The paper argues that this procedure geolocates road users accurately on map data (e.g., Google Maps) across diverse uncalibrated CCTV feeds, and reports competitive 3D detection metrics on the nuScenes validation set while requiring no calibration.

Load-bearing premise

The whole bird's-eye geometry rests on the assumption that a single learned vanishing point, plus two hand-set scalars for depth and sideways offset, fully determines a metric top-down view of an arbitrary street scene, which breaks if the ground is not flat or the camera is rolled relative to the scene.

Editorial extensions

If this is right

  • Any uncalibrated CCTV feed or internet photo can be converted into a vector BEV map of road users, enabling distance and trajectory measurement without per-camera calibration.
  • Road-user positions can be localized on a geographic map using only the vanishing point and two manual scalars, supporting city-scale analyses of traffic and pedestrian exposure.
  • Videos can be reduced to anonymous spatiotemporal token streams (object id, class, 3D box, stationary status), which can be stored and retrieved compactly while preserving privacy.
  • The same framework enables analysis of social distancing violations across hundreds of cameras at city scale, as demonstrated on 857 London CCTV feeds.
  • The reported nuScenes metrics suggest that the method is competitive with camera-only 3D detection methods despite requiring no calibration, implying that the geometric shortcut may be a viable substitute for learned depth in some settings.

Reading between the lines

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

  • The vanishing-point-plus-two-scalars recipe is essentially a degenerate camera model with unknown focal length and height absorbed into two parameters; if the geometry holds, it should extend naturally to calibrated settings by replacing those scalars with true intrinsics and extrinsics, giving a fully parameter-free BEV for any camera.
  • A testable extension is to apply the same homography construction to scenes with visible vertical structures (buildings, poles) by using the learned vanishing point to also constrain scale along the vertical axis, which the current pipeline ignores.
  • The claimed city-scale social-distancing analysis implicitly assumes that the two manual scalars generalize across all cameras in a network; an explicit per-camera error analysis on a held-out set would be a natural next step.
  • The trajectory-based 3D box algorithm only works for moving objects; a direct extension for stationary objects would require inferring orientation from the object's pose relative to the vanishing point, which the paper leaves open.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 6 minor

Summary. The paper presents TopView, a framework for generating vector bird's-eye-view (BEV) representations of road users from uncalibrated street-level images. It combines a learned vanishing-point estimator (MobileNet backbone), YOLOv5m detection with DeepSORT tracking, a heuristic geometric transformation that converts 2D bounding boxes into 3D bounding boxes using the vanishing point and trajectory lines, and a homography module that maps image points to a BEV vector space using four automatically selected points. The authors demonstrate qualitative applications to traffic analysis, geolocalisation on Google Maps, occupancy analysis, and social-distancing measurement. The central claim is that accurate metric BEV geolocalisation is possible without camera intrinsic or extrinsic parameters.

Significance. If the claimed capability were rigorously established, the work would be of practical interest for urban analytics, CCTV processing, and agent-based modelling, since it avoids camera calibration and outputs vector representations amenable to distance measurement. The paper also bundles several useful engineering components, including detection, tracking, and an anonymised token representation of road users. However, the current manuscript does not substantiate the core geometric claims: no quantitative evaluation of BEV accuracy is provided, the geolocalisation validation is qualitative and partly circular, and the nuScenes results in Table 4 are presented without any protocol. The paper includes no code, data, or trained models, limiting reproducibility. On balance, the significance is real but conditional on substantial additional evidence and corrections.

major comments (4)
  1. [§3.1 (Homography) and §4.2 (A simple manual calibration)] The homography construction from a single learned vanishing point and two hand-chosen horizontal lines, followed by a manual z/x calibration, does not determine a metric mapping of the ground plane. For a pinhole camera, a single vanishing point fixes the horizon (and pitch) only under additional assumptions about roll, and it leaves focal length, camera height, and roll unconstrained. Consequently, the projective transformation between the image quadrilateral and the BEV rectangle can have an arbitrary anisotropic scale between the two ground-plane axes, and a global scale (z) and x-shift cannot compensate for this direction-dependent distortion. Therefore the paper does not support its advertised capabilities of metric distance measurement (Application 3) or high-precision GPS-scale geolocalisation (Application 1).
  2. [§4.2 and Figs. 1, 6] The validation of geolocalisation is circular. Section 4.2 describes a manual calibration tool that adjusts the estimated BEV map by two scalar values (z and x) to align it with Google Maps, and the same Google Maps overlay is then presented (Figs. 1 and 6) as qualitative evidence of 'high precision' localisation. No independent ground-truth coordinates, error metrics in metres, or withheld test sites are provided. Thus the claim of high validation in geolocating road users is not established.
  3. [Table 4] The quantitative comparison with nuScenes in Table 4 reports a full set of 3D detection metrics (mATE, mASE, mAOE, mA VE, mAAE, NDS, mAP) for TopView, but the experimental procedure is not described anywhere. It is unclear how the method, which is designed for uncalibrated street-level images, was applied to the nuScenes validation set, which cameras and sequences were used, whether the manual z/x calibration was applied, and what annotations served as ground truth. Without this information, the TopView row is unverifiable and cannot support the claim of competitive 3D detection performance.
  4. [Algorithm 1] Algorithm 1 is under-specified as a method for producing 3D bounding boxes. The algorithm computes an orientation label (e.g., 'turning left' or 'side view') from the intersection of the trajectory line with the top edge of the 2D box, but it does not specify how that label is converted into a concrete 3D bounding box, including its dimensions, the coordinates of its eight corners, or its relationship to the 2D box and the vanishing point. Consequently, the 3D bounding box outputs shown in Fig. 8 are not reproducible from the description.
minor comments (6)
  1. [§3.1] The architecture figure is referenced as 'Fig. ??' in Section 3.1, but no such figure is present; the reader cannot see the intended data flow.
  2. [Throughout] The text consistently uses 'T able' instead of 'Table' (e.g., 'T able 1', 'T able 4'), and the table captions are inconsistently formatted.
  3. [References] References [20] and [21] are both listed as 'Fast R-CNN'; one should refer to the original R-CNN or the duplicate should be removed.
  4. [Fig. 5 caption] The caption contains the stray phrase 'a given image input to a Birds eye view', which appears to be a copy-editing artifact.
  5. [§3.4] The vanishing-point training dataset is described only by source names and image counts; the paper does not explain how ground-truth vanishing points were obtained for the 172,576 images.
  6. [§4.2] The manual calibration tool is not fully specified: the paper does not state how the user chooses the z and x values, whether the tool provides visual feedback, or how sensitive the final BEV localisation is to those values.

Circularity Check

1 steps flagged · score 6.0 of 10

The geolocation validation is partly circular: a two-parameter manual calibration aligned to Google Maps is used as the evidence for the claimed high-precision localisation.

  1. fitted input called prediction [Section 3.2 (Evaluation), Section 4.2 (A simple manual calibration), Application 1]
    "we evaluate the relationship between the different mapped objects after calibrating the image to its geolocation. We used Google Maps as a qualitative measure for verifying the localisations of the objects from the image plane to the real-world coordinate. ... Fig. 9 shows a simple tool for calibrating a birds eye map based on two values: 1) the z-value and 2) the x-value ... this tool provides additional control over the quality of the bird’s eye map for manual calibration when necessary, particularly when linked to a Google Map."

    The claimed high-precision geolocation of road users is not an independent test of the learned vanishing-point model or the homography. The evaluation is explicitly done 'after calibrating the image to its geolocation', and the reference used is the Google Map overlay. The same overlay is the target of the manual calibration tool, which adjusts exactly a z-scale and an x-shift 'when linked to a Google Map'. Once those two parameters are chosen to make the BEV overlay fit the map, reporting the overlay agreement as validation (Application 1: 'We showed a high precision in localising road users in the BEV map when compared to the road layouts of the camera feeds') is circular: the agreement is a consequence of the fitted parameters, not a prediction.

full rationale

The framework contains substantial non-circular components: a supervised vanishing-point regressor trained on external datasets with ground-truth VPs, YOLO/DeepSORT detection and tracking, and a deterministic 2D-to-3D box construction from trajectory lines and the VP. None of these reduce to the target BEV claims by construction. The clearest circularity is in the geolocation validation: the paper calibrates the estimated BEV to Google Maps using a manual z/x tool and then presents the Google Maps overlay as evidence of high-precision localisation. That makes the headline geolocation claim partly fitted rather than independently predicted. The separate geometric concern that a single VP plus two arbitrary horizontal lines cannot determine a metric homography (leaving anisotropic scale and roll ambiguity) is a correctness threat to the metric and distance-measurement claims, but it is not itself a self-referential reduction, so it is not counted as a separate circular step. The nuScenes comparison (Table 4) reports TopView metrics without describing an evaluation protocol, which is an unsupported-evidence concern rather than a demonstrated circularity. The one self-citation ([4], an urban-analytics review) is not load-bearing. Overall, the central BEV rendering pipeline is independent, but the claimed validation of geolocalisation is partially circular, giving a score of 6.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

The central claim relies on a geometric construction that is not fully specified, plus two manual calibration scalars that are fitted to the target map. No new physical entities are introduced.

free parameters (3)
  • z-value (manual BEV calibration) = not reported
    Used in Section 4.2 to adjust the spatial scale of the BEV map to align with Google Maps; chosen by hand per scene, so the resulting localization is partly fitted.
  • x-value (manual BEV calibration) = not reported
    Used alongside z-value to shift road users in the x-coordinate; a hand-chosen free parameter that affects the final BEV geolocalisation.
  • Homography reference line positions = not reported
    The homography is computed from four intersection points of radial lines from the VP with 'upper and lower horizontal lines' whose locations are not specified; different choices change the BEV mapping.
assumptions (4)
  • domain assumption The scene ground is a planar surface, so points on the ground plane can be mapped to a rectangle in a top-down view via homography.
    Invoked in the Homography paragraph of Section 3.1 when constructing the four-point mapping. Real street scenes often have sloped or uneven surfaces, which would break the homography.
  • domain assumption A single predicted vanishing point (with the implied horizon) is sufficient to determine the perspective geometry of an arbitrary uncalibrated camera.
    The VP model outputs only two coordinates; no camera roll, pitch, or focal length is estimated. The homography construction assumes the camera has negligible roll and that the VP lies on the horizon.
  • ad hoc to paper The orientation of a road user can be inferred by intersecting its trajectory line with the top edge of its 2D bounding box and comparing to the VP (Algorithm 1).
    This heuristic is presented in Algorithm 1 with no validation or comparison to learned 3D orientation methods.
  • ad hoc to paper A 3D bounding box can be determined from a 2D box, orientation, and VP without metric scale or camera height.
    The algorithm outputs a 3D box 'based on F and 2D box dimensions' but no metric units or ground plane calibration are provided, so the boxes are pseudo-3D.

how reviews work

0 comments
Cite this review

Pith. "Pith review of TopView: Vectorising road users in a bird's eye view from uncalibrated street-level imagery with deep learning." pith.science (2026). https://pith.science/paper/2DMOXMWB

@misc{pith2026241216229,
  author       = {Pith},
  title        = {Pith review of: TopView: Vectorising road users in a bird's eye view from uncalibrated street-level imagery with deep learning},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/2DMOXMWB}},
  note         = {Machine review of arXiv:2412.16229}
}
read the original abstract

Generating a bird's eye view of road users is beneficial for a variety of applications, including navigation, detecting agent conflicts, and measuring space occupancy, as well as the ability to utilise the metric system to measure distances between different objects. In this research, we introduce a simple approach for estimating a bird's eye view from images without prior knowledge of a given camera's intrinsic and extrinsic parameters. The model is based on the orthogonal projection of objects from various fields of view to a bird's eye view by learning the vanishing point of a given scene. Additionally, we utilised the learned vanishing point alongside the trajectory line to transform the 2D bounding boxes of road users into 3D bounding information. The introduced framework has been applied to several applications to generate a live Map from camera feeds and to analyse social distancing violations at the city scale. The introduced framework shows a high validation in geolocating road users in various uncalibrated cameras. It also paves the way for new adaptations in urban modelling techniques and simulating the built environment accurately, which could benefit Agent-Based Modelling by relying on deep learning and computer vision.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

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

  1. Transforming CCTV cameras into NO$_2$ sensors at city scale for adaptive policymaking

    cs.LG 2024-12 reject novelty 6.0 of 10

    A graph neural network trained on CCTV traffic flows, path signatures, and weather predicts hourly NO2 concentrations across London, but validation is limited and no baseline is reported.

Reference graph

Works this paper leans on

71 extracted references · 48 canonical work pages · cited by 1 Pith paper

  1. [1]

    Ma, Y., et al.: Vision-Centric BEV Perception: A Survey. arXiv. Accessed: Dec. 10, 2022 (2022). http://arxiv.org/abs/2208.02797

    arXiv 2022

  2. [2]

    Li, H., et al.: Delving into the Devils of Birds-eye-view Perception: A Review, Evaluation and Recipe. arXiv. Accessed: Dec. 10, 2022 (2022). http://arxiv.org/ abs/2209.05324

    work page Pith review arXiv 2022

  3. [3]

    Houston, J., et al.: One Thousand and One Hours: Self-driving Motion Prediction Dataset. arXiv. Accessed: Dec. 10, 2022 (2020). http://arxiv.org/abs/2006.14480

    arXiv 2020

  4. [4]

    Cities 96, 102481 (2020) https://doi.org/10.1016/j.cities.2019.102481

    Ibrahim, M.R., Haworth, J., Cheng, T.: Understanding cities with machine eyes: A review of deep computer vision in urban analytics. Cities 96, 102481 (2020) https://doi.org/10.1016/j.cities.2019.102481

    work page arXiv 2020

  5. [5]

    Roddick, T., Kendall, A., Cipolla, R.: Orthographic Feature Transform for Monocular 3D Object Detection. arXiv. Accessed: Dec. 10, 2022 (2018). http: //arxiv.org/abs/1811.08188

    arXiv 2018

  6. [6]

    In: 2010 IEEE International Conference on Intelligent Computing and Intelligent Systems, Xiamen, China, pp

    Xingfang, Y., Yumei, H., Feng, G.: A simple camera calibration method based on sub-pixel corner extraction of the chessboard image. In: 2010 IEEE International Conference on Intelligent Computing and Intelligent Systems, Xiamen, China, pp. 688–692 (2010). https://doi.org/10.1109/ICICISYS.2010.5658280

    work page arXiv 2010

  7. [7]

    IEEE Trans

    Schoepflin, T.N., Dailey, D.J.: Dynamic camera calibration of roadside traffic management cameras for vehicle speed estimation. IEEE Trans. Intell. Transp. Syst. 4(2), 90–98 (2003) https://doi.org/10.1109/TITS.2003.821213

    work page arXiv 2003

  8. [8]

    In: 2019 IEEE 5th International Conference on Wireless and Telematics (ICWT), Yogyakarta, Indonesia, pp

    Mardiati, R., Mulyana, E., Maryono, I., Usman, K., Priatna, T.: The derivation of matrix transformation from pixel coordinates to real-world coordinates for vehicle 22 trajectory tracking. In: 2019 IEEE 5th International Conference on Wireless and Telematics (ICWT), Yogyakarta, Indonesia, pp. 1–5 (2019). https://doi.org/10. 1109/ICWT47785.2019.8978254

    work page arXiv 2019

Show all 71 references
  1. [9]

    International Journal of Scientific & Engineering Research 3(5), 4 (2012)

    Venkatesh, M., Vijayakumar, P.: A simple birds eye view transformation tech- nique. International Journal of Scientific & Engineering Research 3(5), 4 (2012)

    2012

  2. [10]

    Sensors 12(4), 4431–4446 (2012) https://doi.org/10

    Lin, C.-C., Wang, M.-S.: A vision based top-view transformation model for a vehicle parking assistant. Sensors 12(4), 4431–4446 (2012) https://doi.org/10. 3390/s120404431

    2012

  3. [11]

    In: Proceedings of the International Conference Actual Issues of Mechanical Engineering 2017 (AIME 2017), Tomsk, Russia (2017)

    Kholopov, I.S.: Birds eye view transformation technique in photogrammetric problem of object size measuring at low-altitude photography. In: Proceedings of the International Conference Actual Issues of Mechanical Engineering 2017 (AIME 2017), Tomsk, Russia (2017). https://doi....

    2017 doi

  4. [12]

    Regmi, K., Borji, A.: Cross-View Image Synthesis using Conditional GANs. arXiv. Accessed: Dec. 10, 2022 (2018). http://arxiv.org/abs/1803.03396

    2018 arXiv

  5. [13]

    Mani, K., Daga, S., Garg, S., Shankar, N.S., Jatavallabhula, K.M., Krishna, K.M.: MonoLayout: Amodal scene layout from a single image. arXiv. Accessed: Dec. 10, 2022 (2020). http://arxiv.org/abs/2002.08394

    2020 arXiv

  6. [14]

    Proceedings of the IEEE 111(3), 257–276 (2023)

    Zou, Z., Chen, K., Shi, Z., Guo, Y., Ye, J.: Object detection in 20 years: A survey. Proceedings of the IEEE 111(3), 257–276 (2023)

    2023

  7. [15]

    International journal of computer vision 128, 261–318 (2020)

    Liu, L., Ouyang, W., Wang, X., Fieguth, P., Chen, J., Liu, X., Pietik¨ ainen, M.: Deep learning for generic object detection: A survey. International journal of computer vision 128, 261–318 (2020)

    2020

  8. [16]

    IEEE access 7, 128837–128868 (2019)

    Jiao, L., Zhang, F., Liu, F., Yang, S., Li, L., Feng, Z., Qu, R.: A survey of deep learning-based object detection. IEEE access 7, 128837–128868 (2019)

    2019

  9. [17]

    In: 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR’05), vol

    Dalal, N., Triggs, B.: Histograms of oriented gradients for human detection. In: 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR’05), vol. 1, pp. 886–893 (2005). Ieee

    2005

  10. [18]

    Lowe, G.: Sift-the scale invariant feature transform. Int. J 2(91-110), 2 (2004)

    2004

  11. [19]

    IEEE Intelligent Systems and their applications 13(4), 18–28 (1998)

    Hearst, M.A., Dumais, S.T., Osuna, E., Platt, J., Scholkopf, B.: Support vector machines. IEEE Intelligent Systems and their applications 13(4), 18–28 (1998)

    1998

  12. [20]

    In: Proceedings of the IEEE International Conference on Computer Vision, pp

    Girshick, R.: Fast r-cnn. In: Proceedings of the IEEE International Conference on Computer Vision, pp. 1440–1448 (2015)

    2015

  13. [21]

    In: Proceedings of the IEEE International Conference on Computer Vision, pp

    Girshick, R.: Fast r-cnn. In: Proceedings of the IEEE International Conference on Computer Vision, pp. 1440–1448 (2015) 23

    2015

  14. [22]

    In: 2017 IEEE Con- ference on Computer Vision and Pattern Recognition (CVPR), Honolulu, HI, pp

    Redmon, J., Farhadi, A.: Yolo9000: Better, faster, stronger. In: 2017 IEEE Con- ference on Computer Vision and Pattern Recognition (CVPR), Honolulu, HI, pp. 6517–6525 (2017). https://doi.org/10.1109/CVPR.2017.690

    2017 doi

  15. [23]

    https://github.com/ultralytics/yolov5 (2021)

    Ultralytics: YOLOv5. https://github.com/ultralytics/yolov5 (2021)

    2021

  16. [24]

    Redmon, J., Farhadi, A.: YOLOv3: An Incremental Improvement (2018)

    2018

  17. [25]

    In: Computer Vision–ECCV 2016: 14th European Conference, Amsterdam, The Netherlands, October 11–14, 2016, Proceedings, Part I 14, pp

    Liu, W., Anguelov, D., Erhan, D., Szegedy, C., Reed, S., Fu, C.-Y., Berg, A.C.: Ssd: Single shot multibox detector. In: Computer Vision–ECCV 2016: 14th European Conference, Amsterdam, The Netherlands, October 11–14, 2016, Proceedings, Part I 14, pp. 21–37 (2016). Springer

    2016

  18. [26]

    In: Proceedings of the Seventh IEEE International Conference on Computer Vision, Kerkyra, Greece, pp

    Zhang, Z.: Flexible camera calibration by viewing a plane from unknown orienta- tions. In: Proceedings of the Seventh IEEE International Conference on Computer Vision, Kerkyra, Greece, pp. 666–6731 (1999). https://doi.org/10.1109/ICCV. 1999.791289

    1999

  19. [27]

    In: 2019 IEEE/CVF International Conference on Computer Vision (ICCV), Seoul, Korea (South), pp

    Nassar, A., Lefevre, S., Wegner, J.D.: Simultaneous multi-view instance detec- tion with learned geometric soft-constraints. In: 2019 IEEE/CVF International Conference on Computer Vision (ICCV), Seoul, Korea (South), pp. 6558–6567 (2019). https://doi.org/10.1109/ICCV.2019.00666

    2019

  20. [28]

    Sensors 10(3), 2027–2044 (2010) https: //doi.org/10.3390/s100302027

    Escalera, A.D., Armingol, J.M.: Automatic chessboard detection for intrinsic and extrinsic camera parameter calibration. Sensors 10(3), 2027–2044 (2010) https: //doi.org/10.3390/s100302027

    2010 doi

  21. [29]

    Frana, M.B., Piai, J.C.: A new robust algorithmic for multi-camera calibration with a 1d object under general motions without prior knowledge of any camera intrinsic parameter

    Frana, J.A., Stemmer, M.R., M. Frana, M.B., Piai, J.C.: A new robust algorithmic for multi-camera calibration with a 1d object under general motions without prior knowledge of any camera intrinsic parameter. Pattern Recognit. 45(10), 3636–3647 (2012) https://doi.org/10.1016/j....

    2012 doi

  22. [30]

    In: Proceedings 199 IEEE/IEEJ/JSAI International Confer- ence on Intelligent Transportation Systems (Cat

    Ernst, S., Stiller, C., Goldbeck, J., Roessig, C.: Camera calibration for lane and obstacle detection. In: Proceedings 199 IEEE/IEEJ/JSAI International Confer- ence on Intelligent Transportation Systems (Cat. No.99TH8383), Tokyo, Japan, pp. 356–361 (1999). https://doi.org/10.1...

    1999

  23. [31]

    In: 2019 IEEE Intelligent Vehicles Symposium (IV), Paris, France, pp

    Kim, Y., Kum, D.: Deep learning based vehicle position and orientation estima- tion via inverse perspective mapping image. In: 2019 IEEE Intelligent Vehicles Symposium (IV), Paris, France, pp. 317–323 (2019). https://doi.org/10.1109/ IVS.2019.8814050

    2019

  24. [32]

    Abbas, A., Zisserman, A.: A Geometric Approach to Obtain a Birds Eye View from an Image. arXiv. Accessed: Jan. 16, 2023 (2020). http://arxiv.org/abs/1905. 02231

    2020

  25. [33]

    24 arXiv

    Hendy, N., et al.: FISHING Net: Future Inference of Semantic Heatmaps In Grids. 24 arXiv. Accessed: Dec. 10, 2022 (2020). http://arxiv.org/abs/2006.09917

    2020 arXiv

  26. [34]

    Lee, K.-H., et al.: PillarFlow: End-to-end Birds-eye-view Flow Estimation for Autonomous Driving. arXiv. Accessed: Dec. 10, 2022 (2020). http://arxiv.org/ abs/2008.01179

    2020 arXiv

  27. [35]

    Liu, Y., et al.: PETRv2: A Unified Framework for 3D Perception from Multi- Camera Images. arXiv. Accessed: Jan. 16, 2023 (2022). http://arxiv.org/abs/ 2206.01256

    2022 arXiv

  28. [36]

    Li, Y., et al.: BEVDepth: Acquisition of Reliable Depth for Multi-view 3D Object Detection. arXiv. Accessed: Jan. 16, 2023 (2022). http://arxiv.org/abs/2206. 10092

    2022

  29. [37]

    Zhang, Y., et al.: BEVerse: Unified Perception and Prediction in Birds-Eye-View for Vision-Centric Autonomous Driving. arXiv. Accessed: Jan. 16, 2023 (2022). http://arxiv.org/abs/2205.09743

    2022 arXiv

  30. [38]

    Lu, J., Zhou, Z., Zhu, X., Xu, H., Zhang, L.: Learning Ego 3D Representation as Ray Tracing. arXiv. Accessed: Jan. 16, 2023 (2022). http://arxiv.org/abs/2206. 04042

    2022

  31. [39]

    IEEE Robot

    Pan, B., Sun, J., Leung, H.Y.T., Andonian, A., Zhou, B.: Cross-view semantic segmentation for sensing surroundings. IEEE Robot. Autom. Lett. 5(3), 4867– 4873 (2020) https://doi.org/10.1109/LRA.2020.3004325

    2020

  32. [40]

    Peng, L., Chen, Z., Fu, Z., Liang, P., Cheng, E.: BEVSegFormer: Birds Eye View Semantic Segmentation From Arbitrary Camera Rigs. arXiv. Accessed: Dec. 10, 2022 (2022). http://arxiv.org/abs/2203.04050

    2022 arXiv

  33. [41]

    Bartoccioni, F., Zablocki, ., Bursuc, A., Prez, P., Cord, M., Alahari, K.: LaRa: Latents and Rays for Multi-Camera Birds-Eye-View Semantic Segmentation. arXiv. Accessed: Jan. 16, 2023 (2022). http://arxiv.org/abs/2206.13294

    2022 arXiv

  34. [42]

    In: 2021 IEEE/CVF International Conference on Computer Vision (ICCV), Montreal, QC, Canada, pp

    Can, Y.B., Liniger, A., Paudel, D.P., Gool, L.V.: Structured birds-eye-view traffic scene understanding from onboard images. In: 2021 IEEE/CVF International Conference on Computer Vision (ICCV), Montreal, QC, Canada, pp. 15641–15650 (2021). https://doi.org/10.1109/ICCV48922.2021.01537

    2021

  35. [43]

    Zou, J., et al.: HFT: Lifting Perspective Representations via Hybrid Feature Transformation. arXiv. Accessed: Jan. 16, 2023 (2022). http://arxiv.org/abs/ 2204.05068

    2022 arXiv

  36. [44]

    Gong, S., et al.: GitNet: Geometric Prior-based Transformation for Birds-Eye- View Segmentation. arXiv. Accessed: Jan. 16, 2023 (2022). http://arxiv.org/abs/ 2204.07733 25

    2022 arXiv

  37. [45]

    In: ICCV (2021)

    Liu, S., Zhou, Y., Zhao, Y.: Vapid: A rapid vanishing point detector via learned optimizers. In: ICCV (2021)

    2021

  38. [46]

    Zhou, Z., Farhat, F., Wang, J.Z.: Detecting Dominant Vanishing Points in Nat- ural Scenes with Application to Composition-Sensitive Image Retrieval. arXiv. Accessed: Jan. 16, 2023 (2017). http://arxiv.org/abs/1608.04267

    2017 arXiv

  39. [47]

    In: Computer Vision ECCV 2014

    Lin, T.-Y., et al.: Microsoft coco: Common objects in context. In: Computer Vision ECCV 2014. Lecture Notes in Computer Science, vol. 8693, pp. 740–755 (2014). https://doi.org/10.1007/978-3-319-10602-1 48

    2014 doi

  40. [48]

    https://github.com/nwojke/deepsort.git (2019)

    nwojke: DeepSort. https://github.com/nwojke/deepsort.git (2019)

    2019

  41. [49]

    In: 2016 IEEE Int

    Bewley, A., Ge, Z., Ott, L., Ramos, F., Upcroft, B.: Simple online and realtime tracking. In: 2016 IEEE Int. Conf. Image Process. ICIP, pp. 3464–3468 (2016). https://doi.org/10.1109/ICIP.2016.7533003

    2016

  42. [50]

    Brazil, G., Liu, X.: M3D-RPN: Monocular 3D Region Proposal Network for Object Detection. arXiv. Accessed: Jan. 16, 2023 (2019). http://arxiv.org/abs/ 1907.06038

    2019 arXiv

  43. [51]

    Hung, W.-C., Kretzschmar, H., Casser, V., Hwang, J.-J., Anguelov, D.: LET- 3D-AP: Longitudinal Error Tolerant 3D Average Precision for Camera-Only 3D Detection. arXiv. Accessed: Jan. 16, 2023 (2022). http://arxiv.org/abs/2206. 07705

    2022

  44. [52]

    Rukhovich, D., Vorontsova, A., Konushin, A.: ImVoxelNet: Image to Voxels Pro- jection for Monocular and Multi-View General-Purpose 3D Object Detection. arXiv. Accessed: Jan. 16, 2023 (2021). http://arxiv.org/abs/2106.01178

    2021 arXiv

  45. [53]

    Wang, T., Zhu, X., Pang, J., Lin, D. arXiv. Accessed: Jan. 16, 2023 (2021). http://arxiv.org/abs/2104.10956

    2021 arXiv

  46. [54]

    Shi, S., Wang, X., Li, H.: PointRCNN: 3D Object Proposal Generation and Detec- tion from Point Cloud. arXiv. Accessed: Jan. 16, 2023 (2019). http://arxiv.org/ abs/1812.04244

    2019 arXiv

  47. [55]

    Howard, A.G., et al.: MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications. arXiv. Accessed: Nov. 17, 2022 (2017). http://arxiv. org/abs/1704.04861

    2017 arXiv

  48. [56]

    arXiv:1412.6980 [cs]

    Kingma, D.P., Ba, J.: Adam: A Method for Stochastic Optimization. arXiv:1412.6980 [cs]. Accessed: Apr. 23, 2019 (2015). https://hdl.handle.net/ 11245/1.505367

    2015 arXiv

  49. [57]

    In: 2018 IEEE International Conference 26 on Robotics and Automation (ICRA), Brisbane, QLD, pp

    Chang, C.-K., Zhao, J., Itti, L.: Deepvp: Deep learning for vanishing point detec- tion on 1 million street view images. In: 2018 IEEE International Conference 26 on Robotics and Automation (ICRA), Brisbane, QLD, pp. 4496–4503 (2018). https://doi.org/10.1109/ICRA.2018.8460499

    2018

  50. [58]

    Zhou, Y., Qi, H., Huang, J., Ma, Y.: NeurVPS: Neural Vanishing Point Scanning via Conic Convolution. arXiv. Accessed: Jan. 19, 2023 (2021). http://arxiv.org/ abs/1910.06316

    2021 arXiv

  51. [59]

    https://www.tfljamcams.net/ (2021)

    TfL: TfL London cameras. https://www.tfljamcams.net/ (2021)

    2021

  52. [60]

    In: Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision (W ACV) (2021)

    Nabati, R., Qi, H.: Centerfusion: Center-based radar and camera fusion for 3d object detection. In: Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision (W ACV) (2021)

    2021

  53. [61]

    In: 2015 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) (2015)

    Maturana, D., Scherer, S.: Voxelnet: A 3d convolutional neural network for real-time object recognition. In: 2015 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) (2015)

    2015

  54. [62]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2019)

    Lang, A.H., Vora, S., Caesar, H., Zhou, L., Yang, J., Beijbom, O.: Pointpil- lars: Fast encoders for object detection from point clouds. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2019)

    2019

  55. [63]

    In: IEEE Int

    Zhou, X., Wang, D., Krahenbuhl, P.: Objects as points. In: IEEE Int. Conf. Comput. Vis. (2019)

    2019

  56. [64]

    In: Proceedings of IEEE International Conference on Computer Vision (ICCV) (2021)

    Wang, T., Zhu, X., Pang, J., Lin, D.: Fcos3d: Fully convolutional one-stage monocular 3d object detection. In: Proceedings of IEEE International Conference on Computer Vision (ICCV) (2021)

    2021

  57. [65]

    In: Proceedings of the Conference on Robot Learning (CoRL) (2021)

    Wang, Y., Guizilini, V., Zhang, T., Zhao, H., Solomon, J.M.: Detr3d: 3d object detection from multi-view images via 3d-to-2d queries. In: Proceedings of the Conference on Robot Learning (CoRL) (2021)

    2021

  58. [66]

    In: Proceedings of the Conference on Robot Learning (CoRL) (2022)

    Wang, T., Zhu, X., Pang, J., Lin, D.: Pgd: Probabilistic and geometric depth: Detecting objects in perspective. In: Proceedings of the Conference on Robot Learning (CoRL) (2022)

    2022

  59. [67]

    ArXiv preprint arXiv:2203.05625 (2022)

    Liu, Y., Chen, W., Li, Q., Jia, L., Liu, Z., Li, X., Guan, T.: Petr: Position embedding transformation for multi-view 3d object detection. ArXiv preprint arXiv:2203.05625 (2022)

    2022 arXiv

  60. [68]

    ArXiv preprint arXiv:2203.05650 (2022)

    Huang, T., Guo, J., Wang, H., Liao, R., Xie, E., Han, J., Qiao, Y., Liu, W.: Bevdet: High-performance multi-camera 3d object detection in bird-eye-view. ArXiv preprint arXiv:2203.05650 (2022)

    2022 arXiv

  61. [69]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2020)

    Caesar, H., Bankiti, V., Lang, A.H., Vora, S., Liong, V.E., Xu, Q., Krish- nan, A., Pan, Y., Baldan, G., Beijbom, O.: nuscenes: A multimodal dataset for 27 autonomous driving. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2020)

    2020

  62. [70]

    Garnett, N., Cohen, R., Pe’er, T., Lahav, R., Levi, D.: 3D-LaneNet: End-to- End 3D Multiple Lane Detection. arXiv. Accessed: Jan. 16, 2023 (2019). http: //arxiv.org/abs/1811.10203

    2019 arXiv

  63. [71]

    In: Lecture Notes in Computer Science

    Guo, Y., et al.: Gen-lanenet: A generalized and scalable approach for 3d lane detection. In: Lecture Notes in Computer Science. Computer Vision ECCV 2020, vol. 12366, pp. 666–681 (2020). https://doi.org/10.1007/978-3-030-58589-1 40 28

    2020 doi

Pith tools

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