Pith. sign in

REVIEW 4 major objections 4 minor 42 references

An Optimized YOLOv5 Based Approach For Real-time Vehicle Detection At Road Intersections Using Fisheye Cameras

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

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

Pith's one-line read A day/night-specialized pipeline beats a single YOLOv5 detector on fisheye intersection images by 13.7 points of mAP@0.5.

desk verdict A sensible engineering pipeline for fisheye vehicle detection, but the headline 13.7% gain is an arithmetic artifact of comparing different resolutions and test splits. read the letter →

arxiv 2502.04566 v1 pith:FHFDVUWP submitted 2025-02-06 cs.CV

Md. Jahin Alam , Muhammad Zubair Hasan , Md Maisoon Rahman , Md Awsafur Rahman , Najibul Haque Sarker , Shariar Azad , Tasnim Nishat Islam , Bishmoy Paul
show 3 more authors
Tanvir Anjum Barproda Halder Shaikh Anowarul Fattah
This is my paper · ORCID
classification cs.CV
keywords vehicledetectionfisheyecameraYOLOv5day-nightclassificationtransferlearningensembleupsamplingroadintersection
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 a single object-detection network struggles with fisheye intersection images because day and night scenes fail in different ways. It proposes a pipeline that first classifies each frame as day or night, then sends it to a detector trained and specialized for that condition. The pipeline also oversamples difficult images, transfers learning in stages from generic objects to driving footage to fisheye data, adds pseudo-labels from extra fisheye videos, and merges the best trained models. On a real-world fisheye intersection benchmark, the authors report the full system beats a single YOLOv5 detector by 13.7 points of mAP at the 0.5 IoU threshold.

What carries the argument

The load-bearing mechanism is the day/night router: a lightweight CNN classifier with two $3\times3$ convolution layers, LeakyReLU activations, global average pooling, and fully connected layers that labels each fisheye frame as day or night. That label determines which of two YOLOv5 detectors -- one optimized on day fisheye images, one on night fisheye images -- processes the frame. Around the router, the scheme relies on three data-side mechanisms: 10$\times$ upsampling of images whose mAP scores fall below a threshold so the model sees difficult distorted or shadowed vehicles every epoch; multi-stage transfer learning that moves pretrained weights from a generic object dataset to a dashboard-camera driving dataset to the fisheye data; and a pseudo-dataset of 520 frames from additional fisheye junction videos labeled by the model itself. Finally, a selective ensemble takes the best-scoring trained weight sets and merges their bounding-box predictions, keeping non-overlapping boxes and merging overlaps to cut false negatives.

What would settle it

Run the full pipeline and the same YOLOv5 baseline on a held-out set of fisheye intersection footage from a different city, while recording the day/night router's per-frame accuracy; if the mAP advantage disappears on frames where the router errs, the claimed gain rests entirely on a perfect separator.

Watch

Extended reading notes

Core claim

On the paper's own terms, the discovery is that fisheye vehicle detection at junctions is better treated as two problems, day and night, rather than one. A shallow day/night classifier sends each frame to a YOLOv5 detector trained specifically for that period, which lets the night branch specialize on glare, headlight bloom, and low brightness instead of compromising with the day branch. Around that router, the authors stack multi-stage transfer learning from generic object weights to driving-video weights to fisheye weights, 10$\times$ upsampling of the hardest images, a pseudo-labeled fisheye dataset from additional junction videos, and a selective ensemble of the best-scoring models. The reported result is a 13.7-point gain in mAP@0.5 over a single YOLOv5 model, with the day-night classifier reaching 1.0 accuracy on its test images and inference still running in real time because only one detector branch executes per frame.

Load-bearing premise

The whole scheme assumes the day/night classifier never routes an image to the wrong branch; if it errs on a new camera, city, or weather condition, the 13.7-point gain could shrink or even reverse.

Editorial extensions

If this is right

  • Because the separator routes each frame to exactly one detector, the added day/night specialization does not double inference cost, so the pipeline stays real-time.
  • Including the day-only pseudo-dataset improves day-time fisheye scores by roughly one percent while leaving night scores unchanged, showing that pseudo examples shift the model toward day appearance.
  • The best day-time model reaches 0.844 mAP@0.5 at 768$\times$768 input, and higher resolutions consistently raise recall while lowering precision because the model finds vehicles that the ground-truth labels missed.
  • The best night-time result comes from ensembling three models trained at 768$\times$768; adding further models does not improve the score, indicating those weight sets already carried the available information.

Reading between the lines

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

  • A natural extension beyond the paper is to make the router multi-class -- rain, fog, tunnels, and glare are distinct appearance regimes that could each get a specialized branch without changing the inference architecture.
  • The pseudo-labeling step could be closed into a self-training loop: unlabeled fisheye footage from new junctions could be labeled by the current model, filtered for quality, and folded back in to expand coverage of rare vehicle shapes.
  • The paper's own ground-truth analysis implies the measured mAP understates the model, since detections matching unlabeled vehicles are scored as false positives; a corrected label set would change the absolute numbers and could make the reported 13.7-point margin look different.
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 / 4 minor

Summary. The paper proposes a YOLOv5-based vehicle detection pipeline for fisheye road-intersection images. The pipeline includes a lightweight day/night classifier that routes each image to one of two specialized YOLOv5 detectors, upsampling of challenging images, multi-stage transfer learning from COCO and BDD100k, incorporation of a pseudo-labeled fisheye dataset, and selective ensembling of trained weights. The authors report a 13.7% mAP@0.5 improvement over YOLOv5 on the IEEE VIP Cup 2020 dataset, with day and night results presented separately. However, the supporting experiments are component-level: no table evaluates the full pipeline on a single held-out test set, and the headline 13.7% figure is not backed by a matched comparison.

Significance. If the claimed 13.7% improvement were substantiated by a controlled evaluation, the paper would offer a useful engineering recipe for fisheye traffic surveillance, particularly the day/night-specialized detectors and the resolution-dependent behavior documented in Table V. The paper's strengths are its extensive component experiments, explicit recognition of ground-truth noise in Section IV, and the honest report that the pseudo-dataset adds only a small gain. Its limitations are not merely presentational: the central quantitative claim is currently unsupported, and the evaluation protocol does not protect against selection bias. The contributions are incremental rather than methodologically novel, but within a journal scope that accepts solid engineering evaluations, the approach could be publishable after a major corrective revision.

major comments (4)
  1. [Abstract, Tables III and IV] The 13.7% improvement claimed in the Abstract is not supported by a matched experiment. The only comparison that produces this number is the difference between Table IV's Fish Day result at 768 px (0.844 mAP@0.5) and Table III's YOLOv5 baseline at 416 px (0.707 mAP@0.5). These two rows differ in input resolution, training distribution (day-only versus mixed day/night), and possibly model variant; subtracting them is invalid. Table V gives matched-resolution baselines on the mixed set (e.g., YOLOv5x at 768 px: 0.762), but the proposed pipeline is never evaluated on a single held-out test set that includes both day and night images. Please add an end-to-end evaluation of the full pipeline (separator + specialized detectors + ensemble) on the official test split, with the baseline and proposed method matched in resolution, training data, and evaluation protocol.
  2. [Section III-F and Tables IV, VI, VII] All component choices (input resolution, upsampling threshold in Sec. III-F, ensemble composition in Sec. III-G) are made by comparing mAP on the validation portions of the same data that are then used for reporting final numbers. No separate test set is reserved, so the reported scores are subject to selection bias and the uncertainty of the reported gains is unknown. Please split the VIP Cup data into train/validation/test, perform model selection on validation only, and report final numbers on the untouched test set.
  3. [Section IV and Table V] Section IV documents that the ground-truth annotations miss some vehicles, especially in night images, and this is used to explain the AP/AR trend in Table V. However, the magnitude of this label noise is never quantified, and no re-annotation or cleaned-label evaluation is provided. Without this, the absolute mAP values in Tables IV-VII could be either under- or over-estimates depending on how false positives are counted. Please provide a quantitative label-noise analysis (e.g., fraction of images with missing annotations, or a re-annotated sample with inter-annotator agreement) and report sensitivity of the main results to cleaned labels.
  4. [Section III-C] The day/night separator is reported to achieve 1.0 accuracy, but this is on a test set drawn from the same distribution as its training data. Since the entire pipeline routes each test image to exactly one specialized detector, a separator failure on a new camera or weather condition would be load-bearing. Please evaluate the separator on held-out cameras or across different conditions, and, more importantly, report end-to-end performance with the actual separator decisions rather than assuming perfect routing.
minor comments (4)
  1. [Abstract and Section II-C] There are several typos, including 'Furthurmore' in the Abstract and 'fetaures' in Section II-C; please proofread the manuscript.
  2. [Table III] Table III does not state which data partition and split were used for the baseline comparison; please specify the training/test protocol and the exact model variant.
  3. [Section II-I] In Eq. (2), the terms S(.), D(.), and V(.) are not formally defined; please provide precise mathematical definitions or a table of notation.
  4. [References] Reference [20] is a GitHub repository; please cite a versioned release or a peer-reviewed description of YOLOv5, and ensure all references use a consistent format.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the central claim is an empirical benchmark comparison, and no predicted quantity reduces by construction to the paper's own fitted inputs or self-citations.

full rationale

The paper's contribution is a modified training pipeline applied to a public fisheye vehicle-detection benchmark, with performance measured by mAP@0.5 on ground-truth-labeled data. I examined the derivation chain for each claimed improvement. The day/night separator is trained on 500 labeled day and 500 labeled night images and evaluated on the held-out remainder of the same label source; its reported 1.0 accuracy is an independent empirical result and is not defined in terms of the detection metric. The pseudo dataset is generated using the model's own predictions, but it is used only as training augmentation; the reported day-score improvement is measured against ground-truth fish-eye data, so the final metric is not reduced to the model's own outputs. The challenging-image upsampling selects images based on mAP-based ranking and then retrains; this is a potential selection-bias concern, but the paper does not state that the final evaluation set was used for that selection, and the reported scores remain external measurements rather than identities forced by construction. The abstract's 13.7% figure appears to arise from comparing Table IV's 768px day-only result (0.844) with Table III's 416px mixed-data YOLOv5 baseline (0.707); that difference is an unmatched comparison, which is a correctness/statistical issue, not a circular derivation where the prediction equals its input by definition. There are no load-bearing self-citations, no imported uniqueness theorems, and no renaming of a known result as a new derivation. The paper is self-contained against an external benchmark, so no significant circularity is present.

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

The reported performance rests on several unvalidated design choices and domain assumptions, most notably the reliability of the day/night separator and the correctness of self-generated pseudo labels. These are not machine-checked and are not isolated by ablations.

free parameters (4)
  • mAP threshold for challenging image selection = not specified
    Images are ranked by mAP score and a threshold is set to select challenging cases for 10x upsampling (Section II-F). The threshold is chosen post hoc and its value is not reported.
  • Challenging image upsampling factor = 10x
    Selected challenging images are duplicated ten times in the training set (Section II-F). This factor is a hand-chosen hyperparameter with no ablation.
  • Pseudo dataset size = 520 frames
    The pseudo dataset is formed from 520 frames extracted from five YouTube videos (Section II-E). This number is arbitrary and the frames are not released.
  • Final input resolution = 768x768
    The best models use 768x768 input, selected based on the resolution sweep in Table V. The resolution is a hyperparameter that changes mAP substantially.
assumptions (4)
  • domain assumption COCO and BDD100k pretraining transfers useful vehicle features to fisheye intersection images.
    Section II-D assumes multi-stage transfer learning improves convergence and final accuracy, but no controlled comparison isolates the contribution.
  • domain assumption Day and night images are different enough that separate models improve accuracy.
    Section II-C assumes a day/night separator with 1.0 accuracy will route images to the correct model; if the separator errs, performance drops.
  • ad hoc to paper Pseudo labels generated by the model itself are accurate enough to serve as training targets.
    Section II-E creates a pseudo dataset with no manual verification; the model's own detections become ground truth for additional training.
  • domain assumption The provided VIP Cup ground truth is treated as correct for evaluation, even though the paper documents missing labels.
    Section IV notes many vehicles are unlabeled in the test set, which biases mAP; the paper does not re-annotate or account for this in the reported numbers.

how reviews work

0 comments
Cite this review

Pith. "Pith review of An Optimized YOLOv5 Based Approach For Real-time Vehicle Detection At Road Intersections Using Fisheye Cameras." pith.science (2026). https://pith.science/paper/FHFDVUWP

@misc{pith2026250204566,
  author       = {Pith},
  title        = {Pith review of: An Optimized YOLOv5 Based Approach For Real-time Vehicle Detection At Road Intersections Using Fisheye Cameras},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/FHFDVUWP}},
  note         = {Machine review of arXiv:2502.04566}
}
read the original abstract

Real time vehicle detection is a challenging task for urban traffic surveillance. Increase in urbanization leads to increase in accidents and traffic congestion in junction areas resulting in delayed travel time. In order to solve these problems, an intelligent system utilizing automatic detection and tracking system is significant. But this becomes a challenging task at road intersection areas which require a wide range of field view. For this reason, fish eye cameras are widely used in real time vehicle detection purpose to provide large area coverage and 360 degree view at junctions. However, it introduces challenges such as light glare from vehicles and street lights, shadow, non-linear distortion, scaling issues of vehicles and proper localization of small vehicles. To overcome each of these challenges, a modified YOLOv5 object detection scheme is proposed. YOLOv5 is a deep learning oriented convolutional neural network (CNN) based object detection method. The proposed scheme for detecting vehicles in fish-eye images consists of a light-weight day-night CNN classifier so that two different solutions can be implemented to address the day-night detection issues. Furthurmore, challenging instances are upsampled in the dataset for proper localization of vehicles and later on the detection model is ensembled and trained in different combination of vehicle datasets for better generalization, detection and accuracy. For testing, a real world fisheye dataset provided by the Video and Image Processing (VIP) Cup organizer ISSD has been used which includes images from video clips of different fisheye cameras at junction of different cities during day and night time. Experimental results show that our proposed model has outperformed the YOLOv5 model on the dataset by 13.7% mAP @ 0.5.

Figures

Figures reproduced from arXiv: 2502.04566 by the authors.

Figure 1
Figure 1. Graphical Abstract of the proposed scheme: The Challenging subset of images are up-sampled in number and combined with the rest before creating data partitions (1, 2, 3, 4, 5). From this, sequential partitions are defined and each sequence is passed onto the implemented Pipeline which generalizes the weights of the core model. Scores or evaluation metrics produced by each sequence are looked over carefully and the b… view at source ↗
Figure 2
Figure 2. Backbone Architecture of Scheme (a). The provided input image is trained through a sequential pipeline consisting of Focus (b), Conv Blocks (c), BottleNeck CSP (d) and SPP blocks (e); from which three particular feature maps i.e: A, B, C are extracted and passed onto the next stage processing which is the Neck. for night images. Each data sequence found from the data preparation phase is sent through the pipeline ( … view at source ↗
Figure 3
Figure 3. Implemented Core Model The model comprises of three main parts: backbone, neck and head ( [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (8 more)
Figure 4
Figure 4. Figure 4: The Neck Architecture: Path Aggregation Network [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]
Figure 5
Figure 5. Figure 5: Day-Night Separator Model Page 6 of 16 [PITH_FULL_IMAGE:figures/full_fig_p006_5.png]
Figure 6
Figure 6. Figure 6: Day and Night Image Instances of BDD100k Dataset [PITH_FULL_IMAGE:figures/full_fig_p007_6.png]
Figure 7
Figure 7. Figure 7: Preview of Pseudo Dataset F. Challenging Image Up-sampling Imbalanced classes are a common problem in machine learning classification where there are a disproportionate ra￾tio of observations in each class. Most machine learning algorithms work best when the number of …
Figure 8
Figure 8. Figure 8: Fisheye Image Instances from the VIP CUP 2020 [PITH_FULL_IMAGE:figures/full_fig_p010_8.png]
Figure 9
Figure 9. Figure 9: Accuracy and Loss Performances of the Separator [PITH_FULL_IMAGE:figures/full_fig_p011_9.png]
Figure 10
Figure 10. Figure 10: Illustration of improved predictions obtained from including proposed schemes. The introduced training methods reduce [PITH_FULL_IMAGE:figures/full_fig_p014_10.png]
Figure 11
Figure 11. Figure 11: Three instances of ground truth bounding-boxes [PITH_FULL_IMAGE:figures/full_fig_p015_11.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

42 extracted references · 28 canonical work pages

  1. [1]

    Data- driven intelligent transportation systems: A survey,

    J. Zhang, F.-Y . Wang, K. Wang, W.-H. Lin, X. Xu, and C. Chen, “Data- driven intelligent transportation systems: A survey,” IEEE Transactions on Intelligent Transportation Systems , vol. 12, no. 4, 2011

    work page 2011

  2. [2]

    Traffic accident prediction using 3-d model-based vehicle tracking,

    W. Hu, X. Xiao, D. Xie, T. Tan, and S. Maybank, “Traffic accident prediction using 3-d model-based vehicle tracking,” IEEE transactions on vehicular technology , vol. 53, no. 3, pp. 677–694, 2004

    work page 2004

  3. [3]

    Vision-based active safety system for automatic stopping,

    V . Milan ´es, D. F. Llorca, J. Villagr ´a, J. P ´erez, I. Parra, C. Gonz ´alez, and M. A. Sotelo, “Vision-based active safety system for automatic stopping,” Expert Systems with Applications, vol. 39, no. 12, pp. 11 234– 11 242, 2012

    work page 2012

  4. [4]

    Vehicle ego-motion estimation and moving object detection using a monocular camera,

    Y . N. Koichiro Yamaguchi, Takeo Kato, “Vehicle ego-motion estimation and moving object detection using a monocular camera,” January 2006

    work page 2006

  5. [5]

    Vehicle detection and tracking techniques: A concise review

    “Vehicle detection and tracking techniques: A concise review.”

    work page

  6. [6]

    A cnn vehicle recognition algo- rithm based on reinforcement learning error and error-prone samples,

    Z. Xuze, N. Shengsuo, and H. Teng, “A cnn vehicle recognition algo- rithm based on reinforcement learning error and error-prone samples,” in IOP Conference Series: Earth and Environmental Science , vol. 153, no. 3. IOP Publishing, 2018, p. 032052

    work page 2018

  7. [7]

    Using reinforcement learning with partial vehicle detection for intelligent traffic signal control,

    R. Zhang, A. Ishikawa, W. Wang, B. Striner, and O. K. Tonguz, “Using reinforcement learning with partial vehicle detection for intelligent traffic signal control,” IEEE Transactions on Intelligent Transportation Systems, 2020

    work page 2020

  8. [8]

    Yolov4: Optimal speed and accuracy of object detection,

    A. Bochkovskiy, C.-Y . Wang, and H.-Y . M. Liao, “Yolov4: Optimal speed and accuracy of object detection,” 2020

    2020

Show all 42 references
  1. [9]

    Rich feature hierarchies for accurate object detection and semantic segmentation,

    R. Girshick, J. Donahue, T. Darrell, and J. Malik, “Rich feature hierarchies for accurate object detection and semantic segmentation,” in Proceedings of the IEEE conference on computer vision and pattern recognition, 2014, pp. 580–587

    2014

  2. [10]

    Fast r-cnn,

    R. Girshick, “Fast r-cnn,” in Proceedings of the IEEE international conference on computer vision , 2015, pp. 1440–1448

    2015

  3. [11]

    Faster r-cnn: Towards real- time object detection with region proposal networks,

    S. Ren, K. He, R. Girshick, and J. Sun, “Faster r-cnn: Towards real- time object detection with region proposal networks,” arXiv preprint arXiv:1506.01497, 2015

    2015 arXiv

  4. [12]

    Ssd: Single shot multibox detector,

    W. Liu, D. Anguelov, D. Erhan, C. Szegedy, S. Reed, C.-Y . Fu, and A. C. Berg, “Ssd: Single shot multibox detector,” in European conference on computer vision. Springer, 2016, pp. 21–37

    2016

  5. [13]

    Dssd: Deconvolutional single shot detector,

    C.-Y . Fu, W. Liu, A. Ranga, A. Tyagi, and A. C. Berg, “Dssd: Deconvolutional single shot detector,” arXiv preprint arXiv:1701.06659, 2017

    2017 arXiv

  6. [14]

    R-fcn: Object detection via region- based fully convolutional networks,

    J. Dai, Y . Li, K. He, and J. Sun, “R-fcn: Object detection via region- based fully convolutional networks,” arXiv preprint arXiv:1605.06409 , 2016

    2016 arXiv

  7. [15]

    Feature pyramid networks for object detection,

    T.-Y . Lin, P. Doll´ar, R. Girshick, K. He, B. Hariharan, and S. Belongie, “Feature pyramid networks for object detection,” in Proceedings of the IEEE conference on computer vision and pattern recognition , 2017, pp. 2117–2125

    2017

  8. [16]

    Focal loss for dense object detection,

    T.-Y . Lin, P. Goyal, R. Girshick, K. He, and P. Doll ´ar, “Focal loss for dense object detection,” in Proceedings of the IEEE international conference on computer vision , 2017, pp. 2980–2988

    2017

  9. [17]

    Yolov3: An incremental improvement,

    J. Redmon and A. Farhadi, “Yolov3: An incremental improvement,” CoRR, vol. abs/1804.02767, 2018. [Online]. Available: http://arxiv.org/ abs/1804.02767

    2018 arXiv

  10. [18]

    Efficientdet: Scalable and efficient Page 15 of 16 object detection,

    M. Tan, R. Pang, and Q. V . Le, “Efficientdet: Scalable and efficient Page 15 of 16 object detection,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , 2020, pp. 10 781–10 790

    2020

  11. [19]

    Vehicle detection in urban traffic surveillance images based on convolutional neural networks with feature concatenation,

    F. Zhang, C. Li, and F. Yang, “Vehicle detection in urban traffic surveillance images based on convolutional neural networks with feature concatenation,” Sensors, vol. 19, no. 3, 2019. [Online]. Available: https://www.mdpi.com/1424-8220/19/3/594

    2019

  12. [20]

    Yolov5 base repository,

    “Yolov5 base repository,” 2020, [Online]. Available: https://github.com/ ultralytics/yolov5

    2020

  13. [21]

    Spherical formulation of moving object geometric constraints for monocular fisheye cameras,

    I. Letizia Mariotti, Ciaran Hughes Valeo Visions Systems, “Spherical formulation of moving object geometric constraints for monocular fisheye cameras,” p. 8, October 27-30, 2019

    2019

  14. [22]

    Intelligent traffic monitoring and surveillance with multiple cameras,

    A. Koutsia, T. Semertzidis, K. Dimitropoulos, N. Grammalidis, and K. Georgouleas, “Intelligent traffic monitoring and surveillance with multiple cameras,” in 2008 International Workshop on Content-Based Multimedia Indexing. IEEE, 2008, pp. 125–132

    2008

  15. [23]

    Fisheye lens camera based surveillance system for wide field of view monitoring,

    H. Kim, J. Jung, and J. Paik, “Fisheye lens camera based surveillance system for wide field of view monitoring,” Optik, vol. 127, no. 14, pp. 5636–5646, 2016

    2016

  16. [24]

    Real time multi-vehicle tracking and counting at intersections from a fisheye camera,

    W. Wang, T. Gee, J. Price, and H. Qi, “Real time multi-vehicle tracking and counting at intersections from a fisheye camera,” in 2015 IEEE Winter Conference on Applications of Computer Vision . IEEE, 2015, pp. 17–24

    2015

  17. [25]

    360 detection and tracking algorithm of both pedestrian and vehicle using fisheye images,

    M. Bertozzi, L. Castangia, S. Cattani, A. Prioletti, and P. Versari, “360 detection and tracking algorithm of both pedestrian and vehicle using fisheye images,” in 2015 IEEE intelligent vehicles symposium (iv). IEEE, 2015, pp. 132–137

    2015

  18. [26]

    Vehicle detection in close-up range with a fisheye camera,

    H. Tadjine, M. Hess, and M. Montag, “Vehicle detection in close-up range with a fisheye camera,” in 11th FAST-Zero, 2011

    2011

  19. [27]

    A blind-zone detection method using a rear-mounted fisheye camera with combination of vehicle detection methods,

    D. Dooley, B. McGinley, C. Hughes, L. Kilmartin, E. Jones, and M. Glavin, “A blind-zone detection method using a rear-mounted fisheye camera with combination of vehicle detection methods,” IEEE Transactions on Intelligent Transportation Systems , vol. 17, no. 1, pp. 264–278, 2015

    2015

  20. [28]

    Optimization of a cnn-based object detector for fisheye cameras,

    P. Goodarzi, M. Stellmacher, M. Paetzold, A. Hussein, and E. Matthes, “Optimization of a cnn-based object detector for fisheye cameras,” in 2019 IEEE International Conference on Vehicular Electronics and Safety (ICVES). IEEE, 2019, pp. 1–7

    2019

  21. [29]

    Cspnet: A new backbone that can enhance learning capability of cnn,

    C.-Y . Wang, H.-Y . M. Liao, I.-H. Yeh, Y .-H. Wu, P.-Y . Chen, and J.-W. Hsieh, “Cspnet: A new backbone that can enhance learning capability of cnn,” 2019

    2019

  22. [30]

    Spatial pyramid pooling in deep convolutional networks for visual recognition,

    K. He, X. Zhang, S. Ren, and J. Sun, “Spatial pyramid pooling in deep convolutional networks for visual recognition,” Lecture Notes in Computer Science , p. 346–361, 2014. [Online]. Available: http://dx.doi.org/10.1007/978-3-319-10578-9 23

    2014 doi

  23. [31]

    Path aggregation network for instance segmentation,

    S. Liu, L. Qi, H. Qin, J. Shi, and J. Jia, “Path aggregation network for instance segmentation,” 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition , Jun 2018. [Online]. Available: http://dx.doi.org/10.1109/CVPR.2018.00913

    2018

  24. [32]

    Microsoft coco: Common objects in context,

    T.-Y . Lin, M. Maire, S. Belongie, J. Hays, P. Perona, D. Ramanan, P. Doll ´ar, and C. L. Zitnick, “Microsoft coco: Common objects in context,” in European conference on computer vision . Springer, 2014, pp. 740–755

    2014

  25. [33]

    Bdd100k: A diverse driving video database with scalable annotation tooling,

    F. Yu, W. Xian, Y . Chen, F. Liu, M. Liao, V . Madhavan, and T. Darrell, “Bdd100k: A diverse driving video database with scalable annotation tooling,” arXiv preprint arXiv:1805.04687 , vol. 2, no. 5, p. 6, 2018

    2018 arXiv

  26. [34]

    9mp 360-degree camera, fisheye : Interchange in thailand,

    Panasonicsecurity, “9mp 360-degree camera, fisheye : Interchange in thailand,” Sep 2015. [Online]. Available: https://www.youtube.com/ watch?v=NaKaXWR2QYI

    2015

  27. [35]

    9mp 360-degree camera, fisyeye : Intersection in thailand no.2,

    ——, “9mp 360-degree camera, fisyeye : Intersection in thailand no.2,” Sep 2015. [Online]. Available: https://www.youtube.com/watch? v=0FluC dR2RA

    2015

  28. [36]

    Cropping example of 9mp 360-degree camera (road),

    ——, “Cropping example of 9mp 360-degree camera (road),” Sep 2015. [Online]. Available: https://www.youtube.com/watch?v=k5afYMEKm-0

    2015

  29. [37]

    9mp 360-degree camera, fisheye : Road in thailand,

    ——, “9mp 360-degree camera, fisheye : Road in thailand,” Sep 2015. [Online]. Available: https://www.youtube.com/watch?v=ZKum-Sp5xsw

    2015

  30. [38]

    6.6 9mp 360 degree camera, fisyeye intersection in thailand,

    PanasonicBusiness, “6.6 9mp 360 degree camera, fisyeye intersection in thailand,” Nov 2015. [Online]. Available: https://www.youtube.com/ watch?v=jP79Vgyr0ek

    2015

  31. [39]

    Self-training with noisy student improves imagenet classification,

    Q. Xie, M.-T. Luong, E. Hovy, and Q. V . Le, “Self-training with noisy student improves imagenet classification,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , 2020, pp. 10 687–10 698

    2020

  32. [40]

    The pascal visual object classes (voc) challenge,

    M. Everingham, L. Van Gool, C. K. Williams, J. Winn, and A. Zisser- man, “The pascal visual object classes (voc) challenge,” International journal of computer vision , vol. 88, no. 2, pp. 303–338, 2010

    2010

  33. [41]

    Generalized intersection over union: A metric and a loss for bound- ing box regression,

    H. Rezatofighi, N. Tsoi, J. Gwak, A. Sadeghian, I. Reid, and S. Savarese, “Generalized intersection over union: A metric and a loss for bound- ing box regression,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , 2019, pp. 658–666

    2019

  34. [42]

    Distance-iou loss: Faster and better learning for bounding box regression,

    Z. Zheng, P. Wang, W. Liu, J. Li, R. Ye, and D. Ren, “Distance-iou loss: Faster and better learning for bounding box regression,” in Proceedings of the AAAI Conference on Artificial Intelligence , vol. 34, no. 07, 2020, pp. 12 993–13 000. Page 16 of 16

    2020

Pith tools

Reviewed August 8, 2026 · model on record in the stance chip above.