Pith. sign in

REVIEW 3 major objections 4 minor 36 references

Object Detection Approaches to Identifying Hand Images with High Forensic Values

T0 review · 3 major / 4 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 Fine-tuned YOLOv8 models detect hands more accurately than DETR and DETA transformer models across four hand image datasets, and the best model, combined with a bounding-box area threshold, can flag high-forensic-value video frames for…

desk verdict Useful annotation resource and a fair benchmark on human-labeled sets, but the 11k hands evaluation is circular because the 'ground truth' is YOLOv8n's own output. read the letter →

arxiv 2412.16431 v1 pith:ZALZQEHV submitted 2024-12-21 cs.CV cs.AI

Thanh Thi Nguyen , Campbell Wilson , Imad Khan , Janis Dalins This is my paper · ORCID
classification cs.CVcs.AI
keywords handdetectionforensicvalueYOLOv8DETRDETAobjectsemi-automaticannotationvideoframetriage
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 attempts to establish that fine-tuned YOLOv8 models are the most accurate hand detectors among the tested approaches for forensic triage, beating the transformer-based DETR and DETA models on every dataset. To do this, the authors assemble a combined dataset from EgoHands, the 11k hands dataset, and Open Images, and they contribute bounding-box labels for all 11,076 images in the 11k hands dataset using a semi-automatic process that starts from 500 manually drawn boxes. Their best model, YOLOv8x trained on the combined dataset, reaches an average precision of 0.542 on the combined test set and clearly outperforms earlier YOLOv3/YOLOv4 hand detectors. The payoff is practical: when a detected hand fills a large part of an image or video frame, that frame is flagged as having high forensic value, so forensic experts need only review the flagged frames rather than scan entire videos.

What carries the argument

The machinery that carries the comparison is fine-tuned object detection plus a bounding-box-area ranking rule. The detectors are YOLOv8n and YOLOv8x, DETR-ResNet-50, and DETA-Swin-Large, all fine-tuned on the same four training sets and evaluated with the standard AP and AR metrics. For the 11k hands dataset, the authors build labels through a semi-automatic loop: train a YOLOv8n on 500 manually labelled images, use it to predict boxes on the remaining images, add the 1,000 highest-confidence predictions to the training pool, retrain, repeat once, and finally use the model to label the last 8,576 images; those predicted boxes plus the 2,500 seeded boxes become the dataset's ground truth. The forensic application then takes the detector's output boxes, computes hand area as box height times width, and keeps images or frames whose largest hand exceeds a chosen threshold, so the triage decision is a direct function of the detector output.

What would settle it

Manually re-annotate a random sample of roughly 300 images from the 11k hands dataset without reference to the paper's labels, then compute the average precision of the fine-tuned YOLOv8, DETR, and DETA models against these human labels; if the YOLOv8 models' near-perfect scores drop substantially or their ranking against the transformer models changes, the central comparison is not robust to label quality.

Watch

Extended reading notes

Core claim

The central claim is that YOLOv8n and YOLOv8x, fine-tuned for hand detection, outperform DETR-ResNet-50 and DETA-Swin-Large on all four datasets, with the combined-dataset YOLOv8x as the best overall method (AP 0.542). On EgoHands, YOLOv8n and YOLOv8x reach AP 0.768 and 0.766; on Open Images, YOLOv8x reaches 0.417; on 11k hands, both YOLOv8 variants score near 1.0, but that dataset's labels were generated by the same semi-automatic YOLO loop. The paper further claims that selecting frames by the area of the largest detected hand, using a tunable threshold, successfully identifies high-forensic-value frames from videos and reduces the number of frames an expert must inspect, reporting for example 8 of 680 frames flagged in one video and 41 of 497 in another.

Load-bearing premise

The load-bearing premise is that the semi-automatically generated boxes for the 11k hands dataset are valid ground truth, even though most of them were produced by a YOLOv8n model of the same family that is later evaluated against them.

Editorial extensions

If this is right

  • A detector trained on the combined dataset transfers better across hand-image variety: YOLOv8x (combined) scores AP 0.542 on the combined test set versus 0.070 for the same model trained only on EgoHands, showing that training-data breadth matters more than model size alone.
  • For real forensic video review, setting a hand-area threshold (e.g., 30,000 or 50,000 square pixels) reduces a video to a handful of candidate frames: the paper reports 8 of 680 frames for one example video and 41 of 497 for another.
  • YOLOv8 models are sufficient for practical deployment: YOLOv8n, the smallest variant, already achieves AP 0.768 on EgoHands and 0.998 on 11k hands, so forensic triage does not require the largest model.
  • Transformer detectors are not competitive for this task: DETA-Swin-Large, despite 218M parameters, trails YOLOv8n on the combined test set (AP 0.414 vs 0.499) and performs especially poorly on EgoHands (AP 0.248).

Reading between the lines

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

  • A direct extension the authors do not test is using the detected hand crops as inputs to downstream biometric matching (palmprint or vein recognition); the area threshold would then serve as a recall-oriented pre-filter, and the crop quality would matter more than box size.
  • Because 8,576 of 11,076 labels in the 11k dataset were produced by a YOLOv8n model, a fair robustness check would be to have human annotators re-label a random subset; if IoU between human and model boxes is low, the near-perfect AP scores on that dataset are partly self-confirmation.
  • The paper implicitly assumes that a larger hand area equals higher forensic value; in practice, a small hand showing a scar, tattoo, or vein pattern could be more valuable than a large blurred hand. A testable variant would collect expert ratings on a sample of frames and measure how well area-based ranking correlates with those ratings.
  • The threshold used in the example videos differs per video (30,000 vs 50,000 square pixels), so a practical extension would normalise the threshold by image resolution or camera distance; otherwise a global deployment may need per-video calibration.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 4 minor

Summary. The paper compares YOLOv8n, YOLOv8x, DETR-ResNet-50, and DETA-Swin-Large on four hand-image datasets (EgoHands, 11k hands, Open Images, and a combined set), reporting that YOLOv8 variants outperform the transformer-based detectors across all benchmarks. As a contribution, the authors provide bounding-box annotations for the 11k hands dataset using a semi-automatic pipeline that starts with 500 manually labeled images and then iteratively uses a fine-tuned YOLOv8n to label the remaining images. The best model, YOLOv8x trained on the combined dataset, is then applied to video frames to identify frames containing large hand regions as having high forensic value.

Significance. If the results are valid, the paper offers a practical system for forensic triage and a useful resource (publicly available bounding boxes for 11k hands). The comparison across multiple detectors on several datasets is a reasonable engineering study. However, the central performance claim depends critically on the quality of the 11k hands labels, and because those labels were generated by one of the evaluated models (YOLOv8n) in a self-training loop, the near-perfect AP values on that dataset and the combined-dataset rankings are not independent evidence of detection quality. The forensic-value application is also presented with only anecdotal examples and no quantitative evaluation. These issues limit the current support for the abstract's claims.

major comments (3)
  1. [IV.A and V.A.2] The ground-truth labels for the 11k hands dataset are generated in a semi-automatic loop in which YOLOv8n is fine-tuned on 500 manually labeled images, used to predict labels for the rest, and then re-fine-tuned on 1,000 top-confidence predictions plus 500 manual labels; final labels for 8,576 images are YOLOv8n predictions. These model-generated boxes are then treated as ground truth both for training and for the test partition (Table I: 1,994 test images) used in Table IV and in the combined test set (Table VI). Consequently, the reported AP of 0.998/0.999 for YOLOv8n/YOLOv8x on 11k hands measures consistency with YOLOv8n's own output distribution, not independent localization accuracy, and the combined-dataset ranking (YOLOv8x AP 0.542 vs DETR 0.515 vs DETA 0.414) is potentially inflated by label-style agreement. To support the central claim that YOLOv8 models outperform DETR and DETA, the authors should relabel a random subset (e.g., 500-1000 images) of the 11k hands test partition by human annotators, report inter-annotator agreement, and recompute AP on that held-out human-labeled subset; they should also report the manual verification rate for the auto-generated boxes.
  2. [V.A.2, Table IV] The AP-S and AP-M columns in Table IV contain -1.000 for every method on the 11k hands dataset. Negative average precision is not a defined metric; if these entries are intended to indicate that no small or medium objects exist in the 11k hands images, the table should use a standard placeholder such as "—" or "0" with a footnote explaining the absence of those object-size categories. As printed, the values are numerically nonsensical and obscure the interpretation of the otherwise near-perfect results.
  3. [V.B] The forensic-value application is a stated contribution, but the evaluation is anecdotal: the paper reports detected frame counts for two example videos and shows a few stills, but provides no quantitative measure of triage effectiveness, such as precision/recall against a manually labeled set of high-value frames, the number of missed high-value frames, or a comparison with expert manual review time. The claim that the approach "significantly reduc[es] the time required by forensic experts" (abstract and Section VI) is therefore not substantiated. A small user study or an evaluation on a labeled video subset with a defined ground truth for high forensic value is needed to support this claim.
minor comments (4)
  1. [V.A.3] The text states that YOLOv8n and DETR-ResNet-50 "each achiev[e] an AP of 0.355" on Open Images, while the abstract claims YOLOv8 models outperform DETR and DETA on all datasets; since YOLOv8x is the top performer on that dataset, the claim is technically valid, but the wording in Section V.A.3 should be clarified to avoid the impression that every YOLOv8 variant beats every transformer on every metric.
  2. [Table VII] One video name appears as "ff fIRxz2-M" with an extra space, which is likely a typo; the YouTube ID should be verified and the table cleaned.
  3. [References [32]] Reference [32] is cited as a Hugging Face leaderboard space with "and the Hugging Face Team" as co-authors; this is not a standard archival reference and should be replaced with a citable publication or a proper URL with a retrieval date.
  4. [I] The intro mentions ethical considerations and states the work does not collect new data, but no explicit data-availability statement for the combined dataset or the code is provided beyond the 11k hands box link; adding a statement about where the trained models and evaluation scripts can be obtained would improve reproducibility.

Circularity Check

1 steps flagged · score 6.0 of 10

The 11k hands evaluation is partly circular because its ground-truth boxes are generated by a YOLOv8n self-training loop and then used to score YOLOv8 against DETR/DETA.

  1. self definitional [Section IV.A (Used Datasets), semi-automatic annotation procedure for the 11k hands dataset]
    "We then fine-tune the YOLOv8n model on this 2,500-image dataset and use the fine-tuned model to finally predict labels (bounding boxes) for the rest of the images from the 11k hands dataset. These predicted bounding boxes (i.e., for 8,576 images), along with those of the 2,500-image dataset, are considered as ground-truth labels for images in the 11k hands dataset used in this paper."

    The 11k hands ground truth is defined as YOLOv8n's own predictions: 8,576 of 11,076 boxes come from a self-training loop, with no manual verification reported. When YOLOv8n and YOLOv8x are evaluated on the 11k test partition, their near-perfect AP (0.998/0.999) and AR (1.000) therefore measure agreement with boxes produced by the same model family, not independent localization quality. DETR and DETA, which use different matching and refinement behavior, are scored by IoU against this YOLO-generated yardstick. The 11k test set (1,994 images) is also part of the combined test set, so the combined ranking (YOLOv8x AP 0.542 vs DETR 0.515) inherits the same self-definitional element.

full rationale

The central ranking claim has partial circularity. Section IV.A describes a semi-automatic labeling loop in which a YOLOv8n model, initially trained on 500 manually labeled images, is retrained and then used to predict boxes for the remaining 8,576 images of the 11k hands dataset; these predicted boxes are 'considered as ground-truth labels.' The same dataset is then used for training and for the 11k test partition, and it contributes 1,994 of the 7,788 combined test images. The near-perfect AP/AR scores reported on 11k (YOLOv8n 0.998, YOLOv8x 0.999, AR all 1.000) therefore reflect in part how well YOLO-family detectors reproduce the output distribution of a YOLOv8n teacher, not an independent measurement of hand localization quality. DETR/DETA, which use different assignment and box refinement mechanisms, are penalized by IoU against labels that are themselves YOLO predictions. This is a self-definitional element in the evaluation. That said, the paper is not wholly circular. On the externally labeled EgoHands and Open Images datasets, YOLOv8n/x also outperform the transformer detectors (e.g., EgoHands AP 0.768/0.766 vs DETR 0.702 and DETA 0.248; Open Images AP 0.417 for YOLOv8x vs 0.355 for DETR), so the general conclusion that YOLOv8 is strong on these benchmarks is independently supported. The circularity is partial and concentrated in the 11k-based results and the combined score that inherits them. No self-citation or imported uniqueness theorem is involved. Score 6 reflects that one major evaluation component reduces by construction to the method's own outputs, while independent evidence keeps the paper from being entirely circular.

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

The central claim rests on the self-generated 11k labels, an unvalidated forensic-value proxy, and standard transfer-learning assumptions. No fundamentally new entity or theoretical principle is introduced.

free parameters (3)
  • Initial manual label count = 500 images
    Chosen by hand as the seed set for the semi-automatic annotation pipeline in Section IV.A.
  • Top-confidence selection count = 1,000 images
    Chosen by hand in each self-training iteration in Section IV.A.
  • Forensic-value area threshold = 30,000 or 50,000 square pixels depending on video
    Selected per video in Section V.B to decide which frames have high forensic value. No principled criterion is given.
assumptions (4)
  • domain assumption Images or frames containing a substantial hand portion are deemed to have high forensic value.
    Stated in Section V.B; the paper provides no validation that hand area correlates with forensic importance.
  • domain assumption The semi-automatically predicted bounding boxes for the 11k hands dataset are treated as ground truth.
    Section IV.A treats the model-generated boxes and the seeded manual boxes identically as ground truth for training and evaluation.
  • domain assumption Object detection metrics (AP/AR) on the four datasets are a valid proxy for forensic utility.
    Section IV.B adopts standard COCO-style metrics without linking them to forensic outcomes.
  • domain assumption Pre-trained COCO object detection backbones transfer to hand detection after fine-tuning.
    Section III uses pre-trained DETR, DETA, and YOLOv8 checkpoints assuming domain transfer.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Object Detection Approaches to Identifying Hand Images with High Forensic Values." pith.science (2026). https://pith.science/paper/ZALZQEHV

@misc{pith2026241216431,
  author       = {Pith},
  title        = {Pith review of: Object Detection Approaches to Identifying Hand Images with High Forensic Values},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ZALZQEHV}},
  note         = {Machine review of arXiv:2412.16431}
}
read the original abstract

Forensic science plays a crucial role in legal investigations, and the use of advanced technologies, such as object detection based on machine learning methods, can enhance the efficiency and accuracy of forensic analysis. Human hands are unique and can leave distinct patterns, marks, or prints that can be utilized for forensic examinations. This paper compares various machine learning approaches to hand detection and presents the application results of employing the best-performing model to identify images of significant importance in forensic contexts. We fine-tune YOLOv8 and vision transformer-based object detection models on four hand image datasets, including the 11k hands dataset with our own bounding boxes annotated by a semi-automatic approach. Two YOLOv8 variants, i.e., YOLOv8 nano (YOLOv8n) and YOLOv8 extra-large (YOLOv8x), and two vision transformer variants, i.e., DEtection TRansformer (DETR) and Detection Transformers with Assignment (DETA), are employed for the experiments. Experimental results demonstrate that the YOLOv8 models outperform DETR and DETA on all datasets. The experiments also show that YOLOv8 approaches result in superior performance compared with existing hand detection methods, which were based on YOLOv3 and YOLOv4 models. Applications of our fine-tuned YOLOv8 models for identifying hand images (or frames in a video) with high forensic values produce excellent results, significantly reducing the time required by forensic experts. This implies that our approaches can be implemented effectively for real-world applications in forensics or related fields.

Figures

Figures reproduced from arXiv: 2412.16431 by the authors.

Figure 1
Figure 1. Six example images from the 11k hands dataset [16] with our labels, which are bounding boxes in the pink color. These bounding boxes are in [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. An example of a high forensic value frame obtained from the [PITH_FULL_IMAGE:figures/full_fig_p007_2.png] view at source ↗
Figure 3
Figure 3. An example of a high forensic value frame obtained [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

36 extracted references · 30 canonical work pages

  1. [1]

    Forensic examination of the hand,

    L. Hackman and S. Black, “Forensic examination of the hand,” Journal of the Royal Anthropological Institute , vol. 29, pp. 116–131, 2023

    work page 2023

  2. [2]

    Automated latent fingerprint identification system: A review,

    N. Singla, M. Kaur, and S. Sofat, “Automated latent fingerprint identification system: A review,” Forensic Science International, vol. 309, p. 110187, 2020

    work page 2020

  3. [3]

    Palmprints: An introduction,

    A. Badiye, A. Kamble, and N. Kapoor, “Palmprints: An introduction,” in Textbook of Forensic Science. Springer, 2023, pp. 279–293

    work page 2023

  4. [4]

    Hand geometry recognition,

    D. Zhang and V . Kanhangad, “Hand geometry recognition,” in Encyclopedia of Cryptography, Security and Privacy , S. Jajodia, P. Samarati, and M. Yung, Eds. Springer Berlin Heidelberg, 2021, pp. 1–4. [Online]. Available: https://doi.org/10.1007/978-3-642-27739-9 878-2

    doi:10.1007/978-3-642-27739-9 2021

  5. [5]

    On the intra-subject similarity of hand vein patterns in biometric recognition,

    R. S. Kuzu, E. Maiorana, and P. Campisi, “On the intra-subject similarity of hand vein patterns in biometric recognition,” Expert Systems with Applications , vol. 192, p. 116305, 2022

    work page 2022

  6. [6]

    An interdisci- plinary forensic approach for human remains identification and miss- ing persons investigations,

    A. Dahal, D. McNevin, M. Chikhani, and J. Ward, “An interdisci- plinary forensic approach for human remains identification and miss- ing persons investigations,” Wiley Interdisciplinary Reviews: Forensic Science, no. e1484, pp. 1–34, 2023

    work page 2023

  7. [7]

    A formal model for event reconstruction in digital forensic investigation,

    S. Soltani and S. A. H. Seno, “A formal model for event reconstruction in digital forensic investigation,” Digital Investigation , vol. 30, pp. 148–160, 2019

    work page 2019

  8. [8]

    Object detection in 20 years: A survey,

    Z. Zou, K. Chen, Z. Shi, Y . Guo, and J. Ye, “Object detection in 20 years: A survey,” Proceedings of the IEEE , vol. 111, no. 3, pp. 257–276, 2023

    2023

Show all 36 references
  1. [9]

    Convolutional neural network: a review of models, methodologies and applications to object detection,

    A. Dhillon and G. K. Verma, “Convolutional neural network: a review of models, methodologies and applications to object detection,” Progress in Artificial Intelligence , vol. 9, no. 2, pp. 85–112, 2020

    2020

  2. [10]

    Detection and classification of covid-19 by using faster r-cnn and mask r-cnn on ct images,

    M. E. Sahin, H. Ulutas, E. Yuce, and M. F. Erkoc, “Detection and classification of covid-19 by using faster r-cnn and mask r-cnn on ct images,” Neural Computing and Applications , vol. 35, no. 18, pp. 13 597–13 611, 2023

    2023

  3. [11]

    Object detection using yolo: Challenges, architectural successors, datasets and applications,

    T. Diwan, G. Anirudh, and J. V . Tembhurne, “Object detection using yolo: Challenges, architectural successors, datasets and applications,” Multimedia Tools and Applications , vol. 82, no. 6, pp. 9243–9275, 2023

    2023

  4. [12]

    Single shot multibox detector object detection based on attention mechanism and feature fusion,

    X. Wang, K. Li, B. Shi, L. Li, H. Lin, X. Wang, and J. Yang, “Single shot multibox detector object detection based on attention mechanism and feature fusion,” Journal of Electronic Imaging , vol. 32, no. 2, pp. 023 032–023 032, 2023

    2023

  5. [13]

    Recurrent vision transformers for object detection with event cameras,

    M. Gehrig and D. Scaramuzza, “Recurrent vision transformers for object detection with event cameras,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , 2023, pp. 13 884–13 893

    2023

  6. [14]

    End-to-end object detection with transformers,

    N. Carion, F. Massa, G. Synnaeve, N. Usunier, A. Kirillov, and S. Zagoruyko, “End-to-end object detection with transformers,” in European Conference on Computer Vision. Springer, 2020, pp. 213– 229

    2020

  7. [15]

    NMS strikes back,

    J. Ouyang-Zhang, J. H. Cho, X. Zhou, and P. Kr ¨ahenb¨uhl, “NMS strikes back,” arXiv preprint arXiv:2212.06137 , 2022

    2022 arXiv

  8. [16]

    11k hands: Gender recognition and biometric identification using a large dataset of hand images,

    M. Afifi, “11k hands: Gender recognition and biometric identification using a large dataset of hand images,” Multimedia Tools and Applica- tions, vol. 78, pp. 20 835–20 854, 2019

    2019

  9. [17]

    DexYCB: A benchmark for capturing hand grasping of objects,

    Y .-W. Chao, W. Yang, Y . Xiang, P. Molchanov, A. Handa, J. Tremblay, Y . S. Narang, K. Van Wyk, U. Iqbal, S. Birchfield et al., “DexYCB: A benchmark for capturing hand grasping of objects,” in Proceed- ings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition...

    2021

  10. [18]

    Self- supervised 3D hand pose estimation from monocular RGB via con- trastive learning,

    A. Spurr, A. Dahiya, X. Wang, X. Zhang, and O. Hilliges, “Self- supervised 3D hand pose estimation from monocular RGB via con- trastive learning,” in Proceedings of the IEEE/CVF International Conference on Computer Vision , 2021, pp. 11 230–11 239

    2021

  11. [19]

    End-to-end human pose and mesh reconstruction with transformers,

    K. Lin, L. Wang, and Z. Liu, “End-to-end human pose and mesh reconstruction with transformers,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , 2021, pp. 1954–1963

    2021

  12. [20]

    Understanding human hands in contact at internet scale,

    D. Shan, J. Geng, M. Shu, and D. F. Fouhey, “Understanding human hands in contact at internet scale,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , 2020, pp. 9869–9878

    2020

  13. [21]

    Deep learning based person authentication using hand radiographs: A forensic approach,

    S. V . Joshi and R. D. Kanphade, “Deep learning based person authentication using hand radiographs: A forensic approach,” IEEE Access, vol. 8, pp. 95 424–95 434, 2020

    2020

  14. [22]

    Contextual attention for hand detection in the wild,

    S. Narasimhaswamy, Z. Wei, Y . Wang, J. Zhang, and M. Hoai, “Contextual attention for hand detection in the wild,” in Proceedings of the IEEE/CVF International Conference on Computer Vision, 2019, pp. 9567–9576

    2019

  15. [23]

    Mask R-CNN,

    K. He, G. Gkioxari, P. Doll ´ar, and R. Girshick, “Mask R-CNN,” in Proceedings of the IEEE International Conference on Computer Vision, 2017, pp. 2961–2969

    2017

  16. [24]

    Vision-based hand detection and tracking using fusion of kernelized correlation filter and single-shot detection,

    M. N. Haji Mohd, M. S. Mohd Asaari, O. Lay Ping, and B. A. Rosdi, “Vision-based hand detection and tracking using fusion of kernelized correlation filter and single-shot detection,” Applied Sciences, vol. 13, no. 13, p. 7433, 2023

    2023

  17. [25]

    Ultralytics YOLOv8,

    G. Jocher, A. Chaurasia, and J. Qiu, “Ultralytics YOLOv8,” 2023. [Online]. Available: https://github.com/ultralytics/ultralytics

    2023

  18. [26]

    Mi- crosoft COCO: common objects in context,

    T.-Y . Lin, M. Maire, S. Belongie, L. Bourdev, R. Girshick, J. Hays, P. Perona, D. Ramanan, C. L. Zitnick, and P. Doll ´ar, “Mi- crosoft COCO: common objects in context,” https://arxiv.org/abs/1405. 0312v3, February 2015

    2015

  19. [27]

    DETR (end-to-end object detection) model with ResNet-50 backbone,

    Hugging Face, “DETR (end-to-end object detection) model with ResNet-50 backbone,” https://huggingface.co/facebook/detr-resnet-50, November 2023

    2023

  20. [28]

    Deformable DETR: Deformable transformers for end-to-end object detection,

    X. Zhu, W. Su, L. Lu, B. Li, X. Wang, and J. Dai, “Deformable DETR: Deformable transformers for end-to-end object detection,” in International Conference on Learning Representations , 2021. [Online]. Available: https://openreview.net/forum?id=gZ9hCDWe6ke

    2021

  21. [29]

    Lending a hand: Detecting hands and recognizing activities in complex egocentric interactions,

    S. Bambach, S. Lee, D. J. Crandall, and C. Yu, “Lending a hand: Detecting hands and recognizing activities in complex egocentric interactions,” in Proceedings of the IEEE International Conference on Computer Vision , 2015, pp. 1949–1957

    2015

  22. [30]

    The Open Images Dataset V4: Unified image classification, object detec- tion, and visual relationship detection at scale,

    A. Kuznetsova, H. Rom, N. Alldrin, J. Uijlings, I. Krasin, J. Pont- Tuset, S. Kamali, S. Popov, M. Malloci, A. Kolesnikov et al. , “The Open Images Dataset V4: Unified image classification, object detec- tion, and visual relationship detection at scale,” International Journal ...

    1956

  23. [31]

    OpenImages: A public dataset for large-scale multi-label and multi-class image classification

    I. Krasin, T. Duerig, N. Alldrin, V . Ferrari, S. Abu-El-Haija, A. Kuznetsova, H. Rom, J. Uijlings, S. Popov, S. Kamali, M. Malloci, J. Pont-Tuset, A. Veit, S. Belongie, V . Gomes, A. Gupta, C. Sun, G. Chechik, D. Cai, Z. Feng, D. Narayanan, and K. Murphy, “OpenImages: A publi...

    2017

  24. [32]

    Open object de- tection leaderboard on Hugging Face,

    R. Padilla, A. Roberts, and the Hugging Face Team, “Open object de- tection leaderboard on Hugging Face,” https://huggingface.co/spaces/ rafaelpadilla/object detection leaderboard, September 2023

    2023

  25. [33]

    A survey on performance metrics for object-detection algorithms,

    R. Padilla, S. L. Netto, and E. A. Da Silva, “A survey on performance metrics for object-detection algorithms,” in International Conference on Systems, Signals and Image Processing (IWSSIP) . IEEE, 2020, pp. 237–242

    2020

  26. [34]

    A pre-trained YOLO based hand detection network,

    F. Bruggisser, “A pre-trained YOLO based hand detection network,” https://github.com/cansik/yolo-hand-detection, October 2022

    2022

  27. [35]

    Ethical and legal considerations in biometric data usage—Bulgarian perspective,

    J. Deliversky and M. Deliverska, “Ethical and legal considerations in biometric data usage—Bulgarian perspective,” Frontiers in Public Health, vol. 6, no. 25, pp. 1–5, 2018

    2018

  28. [36]

    Biometric privacy protection: What is this thing called privacy?

    E. Mordini, “Biometric privacy protection: What is this thing called privacy?” IET Biometrics, vol. 12, no. 4, pp. 183–193, 2023

    2023

Pith tools

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