Pith. sign in

REVIEW 4 major objections 5 minor 66 references

Efficient Vision-based Vehicle Speed Estimation

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 YOLOv6-based detector with one added 3D-box parameter replaces RetinaNet in the Transform3D pipeline, cutting median speed error to 0.58 km/h from 0.60 km/h while running 5.5 times faster.

desk verdict Solid applied engineering paper; the headline triple claim is not supported by a single configuration, but the systematic accuracy-cost sweep and public code make it worth a serious referee. read the letter →

arxiv 2505.01203 v1 pith:5JEQGEMK submitted 2025-05-02 cs.CV

Andrej Macko , Lukáš Gajdošech , Viktor Kocur This is my paper · ORCID
classification cs.CV MSC 68U10
keywords vehiclespeedestimation3DboundingboxYOLOv6post-trainingquantizationedgecomputingtrafficcameracalibrationvanishingpointobjectdetection
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 accurate vehicle speed measurement from a single traffic camera does not require a heavy detection backbone, and that the bottleneck can be moved to the deployment side. Working from an existing pipeline that turns 2D detections plus vanishing-point geometry into 3D bounding boxes, the authors swap the RetinaNet detector for a YOLOv6 network modified to predict one extra parameter, $c_c$, which locates the top frontal edge of a vehicle's 3D box in a rectified view. On the BrnoCompSpeed test split, the best configuration (a Small model at 640×360 input) reports a median speed error of 0.58 km/h versus the baseline's 0.60 km/h, higher detection precision and recall, and a 5.5 times faster run. The larger point is that smaller, quantized models are the rational choice for real-world speed enforcement, because gains in 2D box localization do not automatically improve the speed measurement.

What carries the argument

The load-bearing object is the perspective-rectified 3D bounding box, parametrized as an ordinary 2D box plus a single scalar $c_c$: after calibration via vanishing points and scale, the image is transformed so that two vanishing-point directions align with image axes, and the 3D box reduces to a 2D box together with the position of its top frontal edge. The paper's modification to YOLOv6 v3.0 is a head that regresses this scalar along with class and box outputs, trained with the MSE term $L_c$ added to the classification and IoU losses. That single extra parameter lets a light anchor-free detector carry the 3D information that the baseline needed a heavier two-stage detector to produce, and it supplies the stable tracking point—the center of the bottom frontal edge—whose per-frame displacements are aggregated into a speed estimate.

What would settle it

Retrain the Small model on the same training splits but with human-verified 3D box labels instead of the mask-plus-calibration pseudo-labels, and measure the median error on split C: if the median error changes materially (by more than about 0.05 km/h), the reported 0.58 km/h result is an artifact of the inherited annotation procedure.

Watch

Extended reading notes

Core claim

The central claim is that the Transform3D speed-estimation pipeline can be made much cheaper without sacrificing accuracy by replacing its detector with an anchor-free YOLOv6 v3.0, adding a small regression head for the perspective-rectified 3D-box parameter $c_c$, and optionally applying post-training INT8 quantization. In this scheme each detection is a 2D box plus one scalar that fixes the top frontal edge; tracking that box's corrected bottom-frontal-edge center across frames and taking the median per-frame displacement yields the speed. The paper measures, on split C of BrnoCompSpeed with the Small model at 640×360 input, a median speed error of 0.58 km/h, mean error 0.81 km/h, detection precision 91.02% and recall 92.16%, compared with the Transform3D baseline's 0.60 km/h, 0.79 km/h, 87.08% and 83.32%, while the FPS benchmarks put the new models well ahead on the Titan V, the only device on which the baseline was benchmarked. It also finds that larger models' better 2D mAP does not translate into better speed accuracy, and that INT8 quantization costs only a couple of hundredths of km/h in median error while sharply increasing throughput.

Load-bearing premise

The paper assumes that the training labels for the 3D box position, which are generated automatically from vehicle masks and camera calibration rather than checked by a person, are accurate enough that a model trained on them gives true speeds.

Editorial extensions

If this is right

  • The same accuracy on BrnoCompSpeed is reachable at a fraction of the compute, so speed-enforcement systems can be deployed on local edge devices rather than sending video to a server.
  • Post-training INT8 quantization is nearly free in this pipeline, making quantized Small or Nano models the practical default over full-precision large models.
  • Input resolution and model size matter less than expected; extra detection precision does not automatically lower the speed error because the estimate is aggregated over many frames.
  • A desktop GPU can process several camera feeds simultaneously, lowering per-road-section hardware cost.
  • The released code and trained models let other groups reproduce the speed-error numbers without retraining.

Reading between the lines

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

  • The paper does not test whether the same head ports to other efficient detectors, but if the $c_c$ parametrization is the source of the gain, the speed-error ranking should transfer across detector families.
  • A design rule the authors leave implicit: INT8 scale calibration should be chosen with the $c_c$ channel in mind, since that single regression output carries the metric information most likely to be hurt by clipping.
  • Because rectification assumes a stable scene geometry, night, rain, or camera vibration should produce systematic bias in the corrected tracking point rather than random noise—a testable failure signature.
  • Selection by speed error rather than mAP suggests a cheaper model-development loop for traffic cameras, where detection checkpoints are judged by final speed accuracy on a small labelled road segment.
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 an efficient variant of the Transform3D pipeline for vehicle speed estimation from traffic cameras. The main change is replacing the RetinaNet detector with a YOLOv6 v3.0 detector extended by an extra head that regresses the cc parameter of the 3D bounding box in the rectified image. The authors train Nano, Small, Medium, and Large variants, some with knowledge distillation, and evaluate FP32, FP16, and INT8 TensorRT versions on the BrnoCompSpeed dataset using the official evaluation tool. They also report an extensive FPS benchmark across six hardware systems, including edge devices. The paper claims that the best model beats the previous state of the art in median speed error (0.58 vs 0.60 km/h), detection precision (91.02% vs 87.08%), and recall (91.14% vs 83.32%) while being 5.5 times faster, and states that code and trained models are publicly available.

Significance. If the claims held as stated, the practical contribution would be solid: the paper would demonstrate that a modern efficient detector combined with post-training quantization can replace the two-stage detector in an existing geometric speed-estimation pipeline with comparable accuracy and substantially lower computational cost. Strengths of the work include evaluation on a public benchmark with the dataset's official evaluation tool, a multi-hardware FPS comparison including edge devices, publicly released code and models, and a systematic comparison over model size, input resolution, and numerical precision. However, the headline claim mixes metrics from different configurations, the accuracy advantage over the baseline is tiny and is reported without statistical support, and the training/test split is described ambiguously. The methodological novelty is limited because the speed-estimation machinery is inherited from Transform3D, but the study is a useful engineering evaluation if the claims are corrected and the split issue is resolved.

major comments (4)
  1. [Abstract / Section 1 / Tables 2 and 4] The abstract's claim that the best model beats Transform3D simultaneously on median error (0.58 vs 0.60), precision (91.02 vs 87.08), recall (91.14 vs 83.32), and speed (5.5x faster) is not supported by any single row of the results. In Table 2, Small F32 at 640x360 has median error 0.58 and precision 91.02 but recall 92.16; Small F32 at 960x540 has recall 91.14 but median error 0.63 and precision 92.11; Small INT8 at 640x360 has precision 91.02 and recall 91.27 but median error 0.60, which ties Transform3D's 0.60. In Table 4, Small INT8 at 640x360 runs at 302 FPS on Titan V versus 62 FPS for Transform3D at the same input size, giving 4.87x rather than 5.5x. The paper should identify the single best configuration and report its metrics together, or revise the abstract so that each claimed number corresponds to a concrete, reproducible configuration.
  2. [Section 4.1, Table 2] The claimed accuracy advantage is a 0.02 km/h difference in median speed error on a single test split, with no confidence intervals, significance test, or per-session breakdown. Because the paper evaluates 18 model/input-size variants and selects the best post hoc, the observed gap could easily arise from noise. Please add bootstrap confidence intervals or a paired statistical test over vehicles or sessions, and state explicitly how the 'best' variant was selected.
  3. [Section 3.3] The training annotations are pseudo-ground-truth 3D boxes generated by combining Mask R-CNN masks with camera calibration data following [27], but the paper does not analyze the accuracy of these annotations or their effect on the learned cc parameter. The final speed evaluation uses LIDAR ground truth, so the headline number is not circular, but a systematic bias in the pseudo-annotations or in the calibration scale could shift absolute speeds and could differentially affect the compared models. Please quantify annotation error or validate on an independent benchmark or via cross-session transfer.
  4. [Section 3.3 vs Section 4.2] The data split is ambiguous and potentially overlapping. Section 3.3 states that the first four sessions of BrnoCompSpeed are used for training and validation, while Section 4.2 identifies the test split as sessions 4-6. If sessions are 1-indexed, session 4 appears in both training and test, which would invalidate the evaluation; if sessions are 0-indexed, the text should say so explicitly. Please clarify the indexing and provide the exact session-to-split mapping used for the reported results.
minor comments (5)
  1. [Section 3.3] The dataset name 'BoxCar116' is inconsistent with the accompanying reference to BoxCars116k; please correct the spelling and use the same name throughout.
  2. [Section 3.4] The sentence 'After the full three epochs we selected a snapshot of the model' should read 'after the full 30 epochs', since the preceding text states that training ran for 30 epochs.
  3. [Table 2] The input-size label '540 x 960' in the Transform3D rows differs from the '960 x 540' notation used elsewhere; please unify the orientation convention.
  4. [Section 4.1] The text mentions that F16 results are provided in the supplementary information, but no F16 accuracy table is discussed in the main text; add a pointer or a brief summary of those results.
  5. [Section 3.3] The phrase 'contains has 21 videos' and the phrase 'to training and validation' are grammatically incorrect and should be revised.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity found: external LIDAR benchmark anchors the speed claim; the Transform3D self-citation is a baseline, not a load-bearing justification.

full rationale

The central speed-estimation claim is anchored to an external benchmark. BrnoCompSpeed [47] provides GPS-synchronized LIDAR ground truth, and the paper reports results using the dataset's official evaluation tool on the held-out test split C. The predicted quantity, the 3D-box offset cc, is an intermediate learned output trained on annotations derived from the authors' earlier Transform3D procedure, but the downstream median speed error is computed from tracked 3D boxes against independent LIDAR speeds on data not used for fitting. No parameter is fitted to the test speed errors and then renamed as a prediction; camera calibration uses the external method [46] and dataset-provided calibration data. The self-citation to [27] is explicit prior-work inheritance: it supplies the baseline architecture, the cc parametrization, and the annotation procedure, but the claimed improvement over [27] is measured by an external evaluation, not by [27]'s own assertions. This is a normal baseline comparison rather than a circular justification. The abstract's internally inconsistent mixture of metrics (0.58 km/h median error from Small F32 at 640x360, 91.14% recall from Small F32 at 960x540, and a 5.5x speedup not matching Table 4's 302/62 ratio) is a correctness or reporting concern, not a circularity, because it does not reduce any equation or fitted parameter to an input. No circular step can be exhibited from the paper's equations or citations.

Assumptions & free parameters 3 free parameters · 5 assumptions · 0 invented entities

The paper introduces no invented physical entities and no parameter-free derivation; the central claim is an empirical system comparison. The assumptions listed are the geometric calibration, annotation-generation, and ground-truth premises inherited from prior work.

free parameters (3)
  • YOLOv6 model weights including cc head = not enumerated (4.85M to 62.71M parameters)
    All reported speed and detection numbers are produced by models trained on BrnoCompSpeed and BoxCars data; these weights are the central fitted quantities.
  • TensorRT INT8 activation scaling factors = not reported (computed by TensorRT calibrator on 1024 training images)
    The quantized variants' accuracy and FPS depend on calibration-derived scaling factors fitted to a subset of the training data.
  • Training hyperparameters (learning rate, momentum, epochs, warmup) = lr=0.01, momentum=0.937, epochs=30, warmup=3
    Chosen by validation; they shape the trained weights but are not presented as derived quantities.
assumptions (5)
  • standard math Perspective projection and vanishing point geometry map the rectified image plane to metric road-plane distances.
    The entire speed-from-3D-box pipeline relies on this geometric model (Section 3.1, Figure 1); it is standard but not re-derived here.
  • domain assumption The camera calibration method of Sochor et al. [46] yields a sufficiently accurate metric scale for the BrnoCompSpeed scenes.
    No recalibration is performed in the paper; calibration errors translate directly into speed errors.
  • domain assumption Pseudo-ground-truth 3D bounding boxes generated from Mask R-CNN masks and calibration data following [27] are accurate enough to train the cc head.
    Systematic errors in these annotations would bias the trained models and could affect absolute speed accuracy even if the relative comparison to [27] is fair.
  • domain assumption BrnoCompSpeed ground-truth speeds from GPS-synchronized LIDAR are correct and the official evaluation tool is unbiased.
    All reported speed errors are measured against this external ground truth (Section 4.1).
  • domain assumption The IOU tracker and median inter-frame distance aggregation produce unbiased speed estimates for the tested traffic scenes.
    Tracking failures or nonlinear vehicle motion between frames would add error that the paper does not model (Section 3.1).

how reviews work

0 comments
Cite this review

Pith. "Pith review of Efficient Vision-based Vehicle Speed Estimation." pith.science (2026). https://pith.science/paper/5JEQGEMK

@misc{pith2026250501203,
  author       = {Pith},
  title        = {Pith review of: Efficient Vision-based Vehicle Speed Estimation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/5JEQGEMK}},
  note         = {Machine review of arXiv:2505.01203}
}
read the original abstract

This paper presents a computationally efficient method for vehicle speed estimation from traffic camera footage. Building upon previous work that utilizes 3D bounding boxes derived from 2D detections and vanishing point geometry, we introduce several improvements to enhance real-time performance. We evaluate our method in several variants on the BrnoCompSpeed dataset in terms of vehicle detection and speed estimation accuracy. Our extensive evaluation across various hardware platforms, including edge devices, demonstrates significant gains in frames per second (FPS) compared to the prior state-of-the-art, while maintaining comparable or improved speed estimation accuracy. We analyze the trade-off between accuracy and computational cost, showing that smaller models utilizing post-training quantization offer the best balance for real-world deployment. Our best performing model beats previous state-of-the-art in terms of median vehicle speed estimation error (0.58 km/h vs. 0.60 km/h), detection precision (91.02% vs 87.08%) and recall (91.14% vs. 83.32%) while also being 5.5 times faster.

Figures

Figures reproduced from arXiv: 2505.01203 by the authors.

Figure 1
Figure 1. When vehicles travel on a straight road their 3D [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Example of mosaic data augmentation with their [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Quantization increases performance of all model [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Comparison of median speed measurement error [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

66 extracted references · 52 canonical work pages

  1. [27]

    Machine Vision and Applications 31(7), 1–15 (2020)

    Kocur, V., Ft´ aˇ cnik, M.: Detection of 3d bounding boxes of vehicles using perspective transformation for accurate speed measurement. Machine Vision and Applications 31(7), 1–15 (2020)

    work page 2020

  2. [1]

    In: 2021 IEEE intelligent vehicles symposium (IV), pp

    Barros, J., Oliveira, L.: Deep speed estimation from syn- thetic and monocular data. In: 2021 IEEE intelligent vehicles symposium (IV), pp. 668–673. IEEE (2021)

    work page 2021

  3. [2]

    Machine Vision and Applications 32(1), 1–13 (2021)

    Bartl, V., ˇSpaˇ nhel, J., Dobeˇ s, P., Jur´ anek, R., Herout, A.: Automatic camera calibration by landmarks on rigid objects. Machine Vision and Applications 32(1), 1–13 (2021)

    work page 2021

  4. [3]

    In: 2022 IEEE International Conference on Systems, Man, and Cybernetics (SMC), pp

    Basterrech, S., Wo´ zniak, M.: Tracking changes using kullback-leibler divergence for the continual learning. In: 2022 IEEE International Conference on Systems, Man, and Cybernetics (SMC), pp. 3279–3285. IEEE (2022)

    work page 2022

  5. [4]

    In: Proceedings of the IEEE/CVF International Conference on Computer Vi- sion, pp

    Bergmann, P., Meinhardt, T., Leal-Taixe, L.: Track- ing without bells and whistles. In: Proceedings of the IEEE/CVF International Conference on Computer Vi- sion, pp. 941–951 (2019)

    work page 2019

  6. [5]

    In: 2016 IEEE interna- tional conference on image processing (ICIP), pp

    Bewley, A., Ge, Z., Ott, L., Ramos, F., Upcroft, B.: Sim- ple online and realtime tracking. In: 2016 IEEE interna- tional conference on image processing (ICIP), pp. 3464–

    work page 2016

  7. [6]

    ACM Transactions on Sensor Net- works (TOSN) 14(3-4), 1–27 (2018)

    Bhardwaj, R., Tummala, G.K., Ramalingam, G., Ram- jee, R., Sinha, P.: Autocalib: automatic traffic camera calibration at scale. ACM Transactions on Sensor Net- works (TOSN) 14(3-4), 1–27 (2018)

    work page 2018

  8. [7]

    In: 2017 14th IEEE international conference on advanced video and signal based surveillance (AVSS), pp

    Bochinski, E., Eiselein, V., Sikora, T.: High-speed tracking-by-detection without using image information. In: 2017 14th IEEE international conference on advanced video and signal based surveillance (AVSS), pp. 1–6. IEEE (2017)

    work page 2017

Show all 66 references
  1. [8]

    arXiv preprint arXiv:2004.10934 (2020)

    Bochkovskiy, A., Wang, C.Y., Liao, H.Y.M.: Yolov4: Op- timal speed and accuracy of object detection. arXiv preprint arXiv:2004.10934 (2020)

    2020 arXiv

  2. [9]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recogni- tion, pp

    Cao, J., Pang, J., Weng, X., Khirodkar, R., Kitani, K.: Observation-centric sort: Rethinking sort for robust multi-object tracking. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recogni- tion, pp. 9686–9696 (2023)

    2023

  3. [10]

    Carion, N., Massa, F., Synnaeve, G., Usunier, N., Kir- illov, A., Zagoruyko, S.: End-to-end object detection with transformers. In: A. Vedaldi, H. Bischof, T. Brox, J.M. Frahm (eds.) Computer Vision – ECCV 2020, pp. 213–

    2020

  4. [11]

    IEEE Transactions on Intelligent Transportation Systems 23(11), 19954–19979 (2022)

    Chen, J., Wang, Q., Cheng, H.H., Peng, W., Xu, W.: A review of vision-based traffic semantic understanding in itss. IEEE Transactions on Intelligent Transportation Systems 23(11), 19954–19979 (2022)

    2022

  5. [12]

    In: European Conference on Computer Vision, pp

    Corral-Soto, E.R., Elder, J.H.: Automatic single-view cal- ibration and rectification from parallel planar curves. In: European Conference on Computer Vision, pp. 813–827. Springer (2014)

    2014

  6. [13]

    In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pp

    Ding, X., Zhang, X., Ma, N., Han, J., Ding, G., Sun, J.: Repvgg: Making vgg-style convnets great again. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pp. 13733–13742 (2021)

    2021

  7. [14]

    IEEE Transactions on Multimedia 25, 8725–8737 (2023)

    Du, Y., Zhao, Z., Song, Y., Zhao, Y., Su, F., Gong, T., Meng, H.: Strongsort: Make deepsort great again. IEEE Transactions on Multimedia 25, 8725–8737 (2023)

    2023

  8. [15]

    In: Proceedings of the British Machine Vision Conference, vol

    Dubsk´ a, M., Herout, A., Sochor, J.: Automatic camera calibration for traffic understanding. In: Proceedings of the British Machine Vision Conference, vol. 4, p. 8. BMVA Press (2014)

    2014

  9. [16]

    Neural Networks 107, 3–11 (2018)

    Elfwing, S., Uchibe, E., Doya, K.: Sigmoid-weighted lin- ear units for neural network function approximation in re- inforcement learning. Neural Networks 107, 3–11 (2018)

    2018

  10. [17]

    In: 2021 IEEE/CVF International Conference on Computer Vi- sion (ICCV), pp

    Feng, C., Zhong, Y., Gao, Y., Scott, M.R., Huang, W.: Tood: Task-aligned one-stage object detection. In: 2021 IEEE/CVF International Conference on Computer Vi- sion (ICCV), pp. 3490–3499. IEEE Computer Society (2021)

    2021

  11. [18]

    In: European Conference on the Applications of Evolutionary Computation, pp

    Filipiak, P., Golenko, B., Dolega, C.: Nsga-ii based auto- calibration of automatic number plate recognition camera for vehicle speed measurement. In: European Conference on the Applications of Evolutionary Computation, pp. 803–818. Springer (2016) 10 Andrej Macko et al

    2016

  12. [19]

    arXiv preprint arXiv:2205.12740 (2022)

    Gevorgyan, Z.: Siou loss: More powerful learn- ing for bounding box regression. arXiv preprint arXiv:2205.12740 (2022)

    2022 arXiv

  13. [20]

    arXiv preprint arXiv:2103.13630 (2021)

    Gholami, A., Kim, S., Dong, Z., Yao, Z., Mahoney, M.W., Keutzer, K.: A survey of quantization methods for efficient neural network inference. arXiv preprint arXiv:2103.13630 (2021)

    2021 arXiv

  14. [21]

    In: Proceedings of the IEEE international conference on computer vision, pp

    He, K., Gkioxari, G., Doll´ ar, P., Girshick, R.: Mask r-cnn. In: Proceedings of the IEEE international conference on computer vision, pp. 2961–2969 (2017)

    2017

  15. [22]

    In: Proceedings of the IEEE conference on computer vision and pattern recognition, pp

    He, K., Zhang, X., Ren, S., Sun, J.: Deep residual learn- ing for image recognition. In: Proceedings of the IEEE conference on computer vision and pattern recognition, pp. 770–778 (2016)

    2016

  16. [23]

    In: Applica- tions of Computer Vision, 2007

    He, X.C., Yung, N.H.: A novel algorithm for estimating vehicle speed from two consecutive images. In: Applica- tions of Computer Vision, 2007. WACV’07. IEEE Work- shop on, pp. 12–12. IEEE (2007)

    2007

  17. [24]

    In: International conference on machine learning, pp

    Ioffe, S., Szegedy, C.: Batch normalization: Accelerat- ing deep network training by reducing internal covariate shift. In: International conference on machine learning, pp. 448–456. pmlr (2015)

    2015

  18. [25]

    Journal of basic Engineering 82(1), 35–45 (1960)

    Kalman, R.E.: A new approach to linear filtering and prediction problems. Journal of basic Engineering 82(1), 35–45 (1960)

    1960

  19. [26]

    In: Proceedings of the 24th Computer Vision Win- ter Workshop, vol

    Kocur, V.: Perspective transformation for accurate de- tection of 3d bounding boxes of vehicles in traffic surveil- lance. In: Proceedings of the 24th Computer Vision Win- ter Workshop, vol. 2, pp. 33–41 (2019)

    2019

  20. [28]

    In: Artificial Neural Networks and Machine Learning – ICANN 2021, pp

    Kocur, V., Ft´ aˇ cnik, M.: Traffic camera calibration via vehicle vanishing point detection. In: Artificial Neural Networks and Machine Learning – ICANN 2021, pp. 628–

    2021

  21. [29]

    2023 IEEE/CVF In- ternational Conference on Computer Vision Workshops (ICCVW) pp

    Lazarevich, I., Grimaldi, M., Kumar, R., Mitra, S., Khan, S., Sah, S.: Yolobench: Benchmarking efficient object detectors on embedded systems. 2023 IEEE/CVF In- ternational Conference on Computer Vision Workshops (ICCVW) pp. 1161–1170 (2023). URL https://api. semanticscholar.o...

    2023

  22. [30]

    0: A full-scale reloading

    Li, C., Li, L., Geng, Y., Jiang, H., Cheng, M., Zhang, B., Ke, Z., Xu, X., Chu, X.: Yolov6 v3. 0: A full-scale reloading. arXiv preprint arXiv:2301.05586 (2023)

    2023 arXiv

  23. [31]

    arXiv preprint arXiv:2209.02976 (2022)

    Li, C., Li, L., Jiang, H., Weng, K., Geng, Y., Li, L., Ke, Z., Li, Q., Cheng, M., Nie, W., et al.: Yolov6: A single-stage object detection framework for industrial applications. arXiv preprint arXiv:2209.02976 (2022)

    2022 arXiv

  24. [32]

    Sensors 23(23), 9527 (2023)

    Li, Y., Zhao, Z., Chen, Y., Zhang, X., Tian, R.: Au- tomatic roadside camera calibration with transformers. Sensors 23(23), 9527 (2023)

    2023

  25. [33]

    IEEE Transactions on Pattern Analy- sis and Machine Intelligence 45(8), 10555–10579 (2023)

    Li, Z., Xu, P., Chang, X., Yang, L., Zhang, Y., Yao, L., Chen, X.: When object detection meets knowledge distil- lation: A survey. IEEE Transactions on Pattern Analy- sis and Machine Intelligence 45(8), 10555–10579 (2023). DOI 10.1109/TPAMI.2023.3257546

    2023

  26. [34]

    In: Proceedings of the IEEE international conference on computer vision, pp

    Lin, T.Y., Goyal, P., Girshick, R., He, K., Doll´ ar, P.: Focal loss for dense object detection. In: Proceedings of the IEEE international conference on computer vision, pp. 2980–2988 (2017)

    2017

  27. [35]

    Applied Sci- ences 14(17) (2024)

    Ling, H., Zhao, T., Zhang, Y., Lei, M.: Engineering ve- hicle detection based on improved yolov6. Applied Sci- ences 14(17) (2024). DOI 10.3390/app14178054. URL https://www.mdpi.com/2076-3417/14/17/8054

    2024 doi

  28. [36]

    In: 2014 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pp

    Luvizon, D.C., Nassu, B.T., Minetto, R.: Vehicle speed estimation by license plate detection and tracking. In: 2014 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pp. 6563–6567. IEEE (2014)

    2014

  29. [37]

    IEEE Transactions on Intelligent Transportation Systems 18(6), 1393–1404 (2017)

    Luvizon, D.C., Nassu, B.T., Minetto, R.: A video-based system for vehicle speed measurement in urban roadways. IEEE Transactions on Intelligent Transportation Systems 18(6), 1393–1404 (2017)

    2017

  30. [38]

    ArXiv abs/2212.07784 (2022)

    Lyu, C., Zhang, W., Huang, H., Zhou, Y., Wang, Y., Liu, Y., Zhang, S., Chen, K.: Rtmdet: An em- pirical study of designing real-time object detectors. ArXiv abs/2212.07784 (2022). URL https://api. semanticscholar.org/CorpusID:254685870

    2022 arXiv

  31. [39]

    In: Image Processing, 2008

    Maduro, C., Batista, K., Peixoto, P., Batista, J.: Estima- tion of vehicle velocity and traffic intensity using recti- fied images. In: Image Processing, 2008. ICIP 2008. 15th IEEE International Conference on, pp. 777–780. IEEE (2008)

    2008

  32. [40]

    In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition Workshops, pp

    Naphade, M., Chang, M.C., Sharma, A., Anastasiu, D.C., Jagarlamudi, V., Chakraborty, P., Huang, T., Wang, S., Liu, M.Y., Chellappa, R., et al.: The 2018 nvidia ai city challenge. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition Workshops, pp. 5...

    2018

  33. [41]

    In: Proceedings of the IEEE conference on computer vision and pattern recognition, pp

    Redmon, J., Divvala, S., Girshick, R., Farhadi, A.: You only look once: Unified, real-time object detection. In: Proceedings of the IEEE conference on computer vision and pattern recognition, pp. 779–788 (2016)

    2016

  34. [42]

    Advances in neural information processing sys- tems 28 (2015)

    Ren, S., He, K., Girshick, R., Sun, J.: Faster r-cnn: To- wards real-time object detection with region proposal networks. Advances in neural information processing sys- tems 28 (2015)

    2015

  35. [43]

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

    Revaud, J., Humenberger, M.: Robust automatic monoc- ular vehicle speed estimation for traffic surveillance. In: Proceedings of the IEEE/CVF International Conference on Computer Vision, pp. 4551–4561 (2021)

    2021

  36. [44]

    In: Pro- ceedings of the IEEE/CVF conference on computer vision and pattern recognition, pp

    Rezatofighi, H., Tsoi, N., Gwak, J., Sadeghian, A., Reid, I., Savarese, S.: Generalized intersection over union: A metric and a loss for bounding box regression. In: Pro- ceedings of the IEEE/CVF conference on computer vision and pattern recognition, pp. 658–666 (2019)

    2019

  37. [45]

    arXiv (2025)

    Sapkota, R., Qureshi, R., Calero, M.F., Badjugar, C., Nepal, U., Poulose, A., Zeno, P., Vaddevolu, U.B.P., Khan, S., Shoman, M., Yan, H., Karkee, M.: Yolo11 to its genesis: A decadal and comprehensive review of the you only look once (yolo) series. arXiv (2025). URL https://ar...

    2025 arXiv

  38. [46]

    Computer Vi- sion and Image Understanding 161, 87–98 (2017)

    Sochor, J., Jur´ anek, R., Herout, A.: Traffic surveillance camera calibration by 3d model bounding box alignment for accurate vehicle speed measurement. Computer Vi- sion and Image Understanding 161, 87–98 (2017)

    2017

  39. [47]

    IEEE Transactions on Intelligent Transportation Systems 20(5), 1633–1643 (2018)

    Sochor, J., Jur´ anek, R., ˇSpaˇ nhel, J., Marˇ s´ ık, L.,ˇSirok` y, A., Herout, A., Zemˇ c´ ık, P.: Comprehensive data set for automatic single camera visual speed measurement. IEEE Transactions on Intelligent Transportation Systems 20(5), 1633–1643 (2018)

    2018

  40. [48]

    IEEE Transactions on In- telligent Transportation Systems PP(99), 1–12 (2018)

    Sochor, J., ˇSpaˇ nhel, J., Herout, A.: Boxcars: Improving fine-grained recognition of vehicles using 3-d bounding boxes in traffic surveillance. IEEE Transactions on In- telligent Transportation Systems PP(99), 1–12 (2018). DOI 10.1109/TITS.2018.2799228

    2018

  41. [49]

    2020 IEEE/CVF Con- ference on Computer Vision and Pattern Recognition (CVPR) pp

    Tan, M., Pang, R., Le, Q.V.: Efficientdet: Scalable and efficient object detection. 2020 IEEE/CVF Con- ference on Computer Vision and Pattern Recognition (CVPR) pp. 10778–10787 (2019). URL https://api. semanticscholar.org/CorpusID:208175544 Efficient Vision-based Vehicle Speed...

    2019

  42. [50]

    Complex & Intelligent Systems 9(4), 4349–4368 (2023)

    Tang, X., Wang, W., Song, H., Zhao, C.: Centerloc3d: monocular 3d vehicle localization network for roadside surveillance cameras. Complex & Intelligent Systems 9(4), 4349–4368 (2023)

    2023

  43. [51]

    Machine Learning and Knowledge Extraction5(4), 1680– 1716 (2023)

    Terven, J., C´ ordova-Esparza, D.M., Romero-Gonz´ alez, J.A.: A comprehensive review of yolo architectures in computer vision: From yolov1 to yolov8 and yolo-nas. Machine Learning and Knowledge Extraction5(4), 1680– 1716 (2023). DOI 10.3390/make5040083. URL https: //www.mdpi.c...

    2023 doi

  44. [52]

    In: Proceedings of the IEEE/CVF international conference on computer vision, pp

    Tian, Z., Shen, C., Chen, H., He, T.: Fcos: Fully convolu- tional one-stage object detection. In: Proceedings of the IEEE/CVF international conference on computer vision, pp. 9627–9636 (2019)

    2019

  45. [53]

    https://github.com/ultralytics/ yolov5 (2020)

    Ultralytics: Yolov5. https://github.com/ultralytics/ yolov5 (2020). Accessed: April 27, 2023

    2020

  46. [54]

    In: Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, pp

    Vuong, K., Tamburo, R., Narasimhan, S.G.: Toward planet-wide traffic camera calibration. In: Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, pp. 8553–8562 (2024)

    2024

  47. [55]

    In: 2017 IEEE international conference on image processing (ICIP), pp

    Wojke, N., Bewley, A., Paulus, D.: Simple online and realtime tracking with a deep association metric. In: 2017 IEEE international conference on image processing (ICIP), pp. 3645–3649. IEEE (2017)

    2017

  48. [56]

    ArXiv abs/2203.16250 (2022)

    Xu, S., Wang, X., Lv, W., Chang, Q., Cui, C., Deng, K., Wang, G., Dang, Q., Wei, S., Du, Y., Lai, B.: Pp-yoloe: An evolved version of yolo. ArXiv abs/2203.16250 (2022)

    2022 arXiv

  49. [57]

    ArXiv abs/2211.15444 (2022)

    Xu, X., Jiang, Y., Chen, W., Huang, Y.L., Zhang, Y., Sun, X.: Damo-yolo : A report on real-time object detection design. ArXiv abs/2211.15444 (2022). URL https://api.semanticscholar.org/ CorpusID:254043744

    2022 arXiv

  50. [58]

    Neurocomputing 204, 222–230 (2016)

    You, X., Zheng, Y.: An accurate and practical calibration method for roadside camera using two vanishing points. Neurocomputing 204, 222–230 (2016)

    2016

  51. [59]

    In: Proceed- ings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp

    Zhang, H., Wang, Y., Dayoub, F., Sunderhauf, N.: Varifo- calnet: An iou-aware dense object detector. In: Proceed- ings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 8514–8523 (2021)

    2021

  52. [60]

    IEEE Trans- actions on Intelligent Transportation Systems 23(9), 14148–14165 (2022)

    Zhang, X., Feng, Y., Angeloudis, P., Demiris, Y.: Monoc- ular visual traffic surveillance: A review. IEEE Trans- actions on Intelligent Transportation Systems 23(9), 14148–14165 (2022)

    2022

  53. [61]

    IEEE transactions on circuits and systems for video technology 23(3), 518–533 (2013)

    Zhang, Z., Tan, T., Huang, K., Wang, Y.: Practical camera calibration from moving objects for traffic scene surveillance. IEEE transactions on circuits and systems for video technology 23(3), 518–533 (2013)

    2013

  54. [62]

    Applied Sciences 11(20), 9680 (2021)

    Zhou, X., Ke, R., Yang, H., Liu, C.: When intelligent transportation systems sensing meets edge computing: Vision and challenges. Applied Sciences 11(20), 9680 (2021)

    2021

  55. [63]

    In: European conference on computer vision, pp

    Zhou, X., Koltun, V., Kr¨ ahenb¨ uhl, P.: Tracking objects as points. In: European conference on computer vision, pp. 474–490. Springer (2020)

    2020

  56. [64]

    ArXiv abs/1904.07850 (2019)

    Zhou, X., Wang, D., Kr¨ ahenb¨ uhl, P.: Objects as points. ArXiv abs/1904.07850 (2019). URL https://api. semanticscholar.org/CorpusID:118714035

    2019 arXiv

  57. [229]

    Springer International Publishing, Cham (2020)

    2020

  58. [639]

    Springer International Publishing, Cham (2021)

    2021

Pith tools

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