Pith. sign in

REVIEW 4 major objections 5 minor 71 references

YOLO11-JDE: Fast and Accurate Multi-Object Tracking with Self-Supervised Re-ID

T0 review · 4 major / 5 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read A lightweight tracker learns to re-identify people without identity labels by exploiting Mosaic-augmented training images.

desk verdict A solid, honest engineering paper whose speed claim outruns its measurement: the FPS table needs a controlled protocol before the central advantage is believable. read the letter →

arxiv 2501.13710 v1 pith:7VH5TISC submitted 2025-01-23 cs.CV cs.AI

Iñaki Erregue , Kamal Nasrollahi , Sergio Escalera This is my paper · ORCID
classification cs.CVcs.AI
keywords multi-objecttrackingjointdetectionandembeddingself-supervisedre-identificationtripletlossMosaicaugmentationYOLO11real-timedataassociation
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper claims that multi-object tracking can run in real time on a small, under-10-million-parameter network while still matching the accuracy of much larger joint-detection-and-embedding (JDE) systems. The key move is training the re-identification branch without any identity annotations: Mosaic augmentation stitches several images into one, showing the same person under different transformations within a single training batch. A triplet loss with hard-positive and semi-hard-negative mining then learns appearance embeddings from those repeated views. The resulting tracker, YOLO11-JDE, reports competitive HOTA, MOTA, and IDF1 scores on MOT17 and MOT20 while running at higher FPS and using up to ten times fewer parameters than earlier JDE methods. If these speed and efficiency numbers hold under matched conditions, appearance-based tracking becomes viable on devices where large detectors cannot run.

What carries the argument

The load-bearing component is the Re-ID branch appended to the YOLO11s decoupled head: two 3x3 convolutional layers with batch norm and SiLU activation followed by a 1x1 convolution that maps to the final embedding dimension (128 in the best configuration). This branch is trained with the triplet loss under hard-positive/semi-hard-negative mining, and the self-supervised signal comes from Mosaic augmentation, which places multiple transformed copies of the same identity in one training image; a custom two-stage association tracker then fuses Mahalanobis distance, cosine embedding distance, IoU, and detection confidence in the final cost matrix.

What would settle it

Run YOLO11-JDE and a competing JDE method (e.g., FairMOT or CSTrack) on the same GPU with the same input resolution, batch size, and detection confidence thresholds, and compare end-to-end FPS; if the speed gap narrows or reverses, the paper's headline efficiency claim fails. Additionally, count the model's learnable parameters after exporting; if the deployed model exceeds 10 million, the size claim is also contradicted.

Watch

Extended reading notes

Core claim

The paper's central claim is that a YOLO11s detector augmented with a dedicated Re-ID branch can perform joint detection and embedding in a fully self-supervised fashion, eliminating the need for identity-labeled tracking datasets. Using Mosaic data augmentation to generate multiple views of each identity in the same image, the model is trained simultaneously for detection and for a triplet loss that pulls same-identity embeddings together and pushes different-identity embeddings apart. This produces discriminative appearance features that, combined with motion, location, and confidence cues in a custom association tracker, yield competitive performance on MOT17 and MOT20 while achieving higher FPS and a parameter count an order of magnitude smaller than top JDE competitors.

Load-bearing premise

The central speed advantage rests on the assumption that the FPS numbers from different papers were measured under comparable hardware, framework, and inference settings, but the paper does not report those conditions, so the tenfold speed gap could be an artifact of the test environment rather than a property of the model.

Editorial extensions

If this is right

  • If the reported efficiency is real, JDE-based MOT can run at interactive frame rates on hardware with limited memory, widening the deployable range to edge devices and embedded cameras.
  • The finding that identity supervision did not improve HOTA or IDF1 suggests that cheap unlabeled detection data from crowd-heavy datasets may be sufficient to learn Re-ID features for tracking.
  • The custom association tracker, which combines motion, appearance, IoU, and confidence, recovers large gains over the default FairMOT tracker on the MOT17 validation set, indicating that tracker hyperparameters must be tuned to the model's specific output distribution.
  • Because the model generalizes to crowded MOT20 scenes without ever training on MOT20, the combination of CrowdHuman data and Mosaic augmentation appears to confer robustness to occlusion and dense crowds.
  • The framework is compatible with semi-supervised training, so a small amount of identity-labeled data can be added to fine-tune or repair specific failure modes without redesigning the training pipeline.

Reading between the lines

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

  • The paper does not quantify energy consumption or latency on embedded hardware, but a tenfold parameter reduction likely translates to proportionally lower inference cost; measuring this on a mobile GPU would be a natural extension.
  • The ablation results imply that detection quality, not Re-ID discriminability, is the current bottleneck for the method; strengthening the detector or decoupling the two tasks, as the authors suggest as future work, could push YOLO11-JDE toward state-of-the-art accuracy.
  • The self-supervised Mosaic-plus-triplet recipe may transfer to other lightweight detectors and other dense-object domains (e.g., retail shelf items, animal counts); the paper only demonstrates it with one backbone and one object class.
  • A stronger test of the self-supervised mechanism would be to ablate Mosaic specifically and measure embedding quality; if identity views from Mosaic are essential, removing it should collapse the Re-ID signal, but the paper's ablations do not isolate this factor directly.
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 / 5 minor

Summary. The paper introduces YOLO11-JDE, a joint detection and embedding (JDE) tracker built on YOLO11s. A Re-ID branch is added to the detection head and trained with a triplet loss using hard-positive and semi-hard-negative mining, in a setting the authors call fully self-supervised and based on Mosaic data augmentation. A custom online data association algorithm combines motion (Kalman filter), appearance (cosine distance of Re-ID embeddings), and location (IoU) cues. The method is evaluated on MOT17 and MOT20 under the private-detection protocol, reporting HOTA/MOTA/IDF1/FPS. The central claims are that YOLO11-JDE achieves competitive tracking accuracy while surpassing existing JDE methods in FPS and using up to ten times fewer parameters.

Significance. If the speed and parameter-efficiency claims hold under a controlled comparison, the method would be a practically useful lightweight real-time tracker that does not require identity-labeled training data. The paper contributes public code, a clear architectural modification of YOLO11s, and an ablation study covering mining strategies, loss margin, confidence filtering, loss weight, embedding dimension, and training data. The self-supervised formulation is interesting because it eliminates identity supervision, and the MOT20 results suggest robustness in crowded scenes. However, the current evidence for the speed superiority is based on an uncontrolled FPS table, and the self-supervised positive-pair mechanism is not sufficiently specified to be reproducible.

major comments (4)
  1. [Section 4.5, Table 8] The FPS comparison is not controlled and therefore cannot support the central claim of surpassing existing JDE methods in speed. The paper does not specify the GPU model, batch size, framework version, TensorRT/FP16 usage, the input resolution used for each baseline, or whether the reported FPS includes the full tracking pipeline (detector + association) or only the detector. Baselines such as FairMOT and CSTrack were originally evaluated at different resolutions (e.g., 1088x608) and on different hardware, while YOLO11-JDE is run at 1280 pixels. As a result, the 35.9 FPS on MOT17 could be a hardware artifact rather than an architectural property. Please provide a controlled benchmark on a single GPU with the same inference settings, or at least a detailed table of hardware and configuration for every method.
  2. [Section 3.2 and 4.2] The self-supervised training mechanism is underspecified and, as written, does not explain how positive pairs arise for the triplet loss when identity annotations are absent. Mosaic augmentation stitches four different images, so a given identity appears only once per Mosaic image; the paper claims that Mosaic 'enables the model to review the same identities under diverse transformations,' but it does not describe any mechanism that creates multiple views of the same identity within an input image or batch. Unless the same source image is intentionally repeated with different augmentations in a batch, the triplet loss would have no valid positive pairs and would collapse. The statement in Section 4.2 that identity annotations are 'generated synthetically if not available' is insufficient; the authors must specify how synthetic identities are assigned, how positive pairs are sampled, and how the label-preservation under Mosaic is implemented.
  3. [Section 4.3, Tables 1-6] The ablations are performed sequentially on a single validation split with no repeated runs or error bars, and many of the reported differences are small enough to be within run-to-run noise (e.g., Table 2: HOTA 56.37 vs 56.03 for margins 0.075 vs 0.025; Table 5: HOTA 56.37 vs 56.27 for dimensions 128 vs 64). The ablations also use a different training budget (30 epochs, 640px) and a different tracker (default FairMOT) than the final model (100 epochs, 1280px, custom tracker), so it is not clear that the selected hyperparameters are optimal for the final inference setup. Please report variance across seeds and, if possible, validate the chosen configuration under the final training and tracking conditions.
  4. [Section 4.5 and Abstract] The descriptor 'competitive results' is not supported when compared against the cited state of the art: on MOT17, YOLO11-JDE achieves HOTA 56.6, whereas FairMOT, CSTrack, TCBTrack, QDTrack, and CountingMOT report HOTA 59.3, 59.3, 62.1, 63.5, and 63.6 respectively. The paper attributes the gap to detection quality, which is acceptable, but then the abstract and introduction overstate the accuracy result. Please either qualify the accuracy claim (e.g., 'competitive among lightweight/self-supervised JDE methods') or provide additional evidence that the accuracy is competitive in a meaningful sense.
minor comments (5)
  1. [Equation (1)] The notation [x]_+ is not defined; please define it as max(0, x) or use the hinge form explicitly.
  2. [Section 3.4] There is a typo: 'Normalized Re-ID embeddings are use to compute' should be 'are used to compute'.
  3. [Section 4.2] Please state explicitly which training data and annotation setting (fully self-supervised vs semi-supervised) is used for the final model whose results appear in Table 8, since the ablation table distinguishes CH, MOT17* from CH, MOT17 but the main experiment does not repeat this distinction.
  4. [Section 4.5] The statement 'the only fully self-supervised method in the comparison' needs justification, as QDTrack also uses self-supervised contrastive learning; if the distinction is that QDTrack uses detection labels or additional supervision, please clarify what exactly makes YOLO11-JDE 'fully self-supervised'.
  5. [Figure 1] The bubble size represents ID switches, but there is no legend explaining the scale; adding a legend or explicit values would improve readability.

Circularity Check

0 steps flagged · score 0.0 of 10

No circular derivation: the paper is an empirical architecture-and-training contribution evaluated on external MOT benchmarks, with no load-bearing self-citation or fitted-input-as-prediction step.

full rationale

The paper contains no claimed derivation that reduces to its own inputs. The central contribution is a modified YOLO11s detector with an added Re-ID branch, trained with triplet loss under a Mosaic-augmentation-based self-supervised identity-assignment scheme, and evaluated on the external MOT17 and MOT20 test sets under the private-detection protocol. The triplet-loss margin, mining strategy, loss weight, and feature dimension are tuned on validation splits (Tables 1-5), which is standard empirical practice rather than circular reasoning: the reported test metrics are not themselves used to fit any parameter. The 'synthetic identities' produced by Mosaic augmentation are a training signal, not a claimed output that is later measured as a prediction; the Re-ID quality in ablations is checked against MOT17 ground-truth identity labels, which are independent of the synthetic instance-level labels used for the self-supervised training runs. No load-bearing step is justified by a self-citation: the authors cite prior work for standard components (triplet loss, Mosaic, FairMOT tracker design), but none of those citations is from the present authors, and none is invoked as an unverified uniqueness theorem or as the sole support for the main claim. The strongest claim, regarding FPS superiority over existing JDE methods, rests on Table 8, which lacks a controlled hardware/framework protocol; however, an uncontrolled speed comparison is a measurement-fairness or correctness concern, not a circularity concern, because the FPS numbers are not derived from the paper's own fitted parameters or definitions. The evaluation is self-contained against external benchmarks, and the paper's own ablations show that the model does not exceed its training signal: it achieves lower HOTA than heavier supervised JDE models, which is consistent with an honest empirical comparison. Thus no pattern of self-definition, fitted-input-called-prediction, self-citation load-bearing, imported uniqueness, ansatz-smuggling, or renaming applies.

Assumptions & free parameters 8 free parameters · 3 assumptions · 0 invented entities

The paper introduces no new physical or conceptual entities. Synthetic identities are temporary training labels generated by Mosaic augmentation, not entities with external existence. The free parameters are all standard empirical hyperparameters tuned on validation data.

free parameters (8)
  • triplet loss margin = 0.075
    Tuned on MOT17 validation around the default m=0.05 (Table 2); directly controls how far apart embeddings must be.
  • triplet loss weight = 1.0
    Ablated in Table 4; balances the Re-ID loss against the detection losses in joint training.
  • embedding dimension = 128
    Ablated in Table 5 among 64, 128, and 256; sets the size of the appearance feature vector.
  • mining strategy = hard positive, semi-hard negative
    Ablated in Table 1; determines which triplets are used in the loss and is claimed to be the best configuration.
  • confidence filtering threshold = 100% (no filtering)
    Ablated in Table 3; using all predictions outperformed filtering to the top 75% or 50%.
  • input resolution = 1280 pixels at inference, 640 or 1280 during training
    Resolution changes the accuracy/speed tradeoff; the ablation setup uses 640 pixels while the final model uses 1280.
  • data association fusion weights = not reported numerically
    The custom cost matrix combines motion, appearance, and IoU; the weights are tuned on the MOT17 training split (Table 7) but not listed.
  • tracklet buffer = 30 frames
    Unmatched tracklets persist for 30 frames to handle occlusions, following FairMOT; the value is not ablated.
assumptions (3)
  • domain assumption Motion between consecutive frames is approximately linear for short tracklets.
    Section 3.4 uses a Kalman filter with linear motion dynamics to predict track locations; strongly nonlinear motion would weaken the motion cue.
  • domain assumption Mosaic augmentation preserves object identity across the four image patches, so same-object patches form valid positive pairs.
    Section 3.2 relies on this to generate synthetic identities for self-supervised Re-ID training.
  • domain assumption Bounding boxes from CrowdHuman and MOT17 training splits provide sufficient foreground supervision for simultaneous detection training.
    Section 3.3 assigns ground-truth identity labels only to foreground predictions; if detection boxes are noisy, triplet mining inherits that noise.

how reviews work

0 comments
Cite this review

Pith. "Pith review of YOLO11-JDE: Fast and Accurate Multi-Object Tracking with Self-Supervised Re-ID." pith.science (2026). https://pith.science/paper/7VH5TISC

@misc{pith2026250113710,
  author       = {Pith},
  title        = {Pith review of: YOLO11-JDE: Fast and Accurate Multi-Object Tracking with Self-Supervised Re-ID},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/7VH5TISC}},
  note         = {Machine review of arXiv:2501.13710}
}
read the original abstract

We introduce YOLO11-JDE, a fast and accurate multi-object tracking (MOT) solution that combines real-time object detection with self-supervised Re-Identification (Re-ID). By incorporating a dedicated Re-ID branch into YOLO11s, our model performs Joint Detection and Embedding (JDE), generating appearance features for each detection. The Re-ID branch is trained in a fully self-supervised setting while simultaneously training for detection, eliminating the need for costly identity-labeled datasets. The triplet loss, with hard positive and semi-hard negative mining strategies, is used for learning discriminative embeddings. Data association is enhanced with a custom tracking implementation that successfully integrates motion, appearance, and location cues. YOLO11-JDE achieves competitive results on MOT17 and MOT20 benchmarks, surpassing existing JDE methods in terms of FPS and using up to ten times fewer parameters. Thus, making our method a highly attractive solution for real-world applications.

Figures

Figures reproduced from arXiv: 2501.13710 by the authors.

Figure 1
Figure 1. Comparative analysis of JDE models on the MOT20 test [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Comparison of tracking architectures: (a) Separate Detection and Embedding (SDE), where detection and Re-ID are performed [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗
Figure 3
Figure 3. Example of four training images using Mosaic data [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Example of consistent identity maintenance across frames on the MOT20-04 sequence despite multiple occlusions (ID 248). [PITH_FULL_IMAGE:figures/full_fig_p008_4.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

71 extracted references · 61 canonical work pages

  1. [1]

    Bot-sort: Robust associations multi-pedestrian tracking, 2022

    Nir Aharon, Roy Orfaig, and Ben-Zion Bobrovsky. Bot-sort: Robust associations multi-pedestrian tracking, 2022. 1, 3

    work page 2022

  2. [2]

    Towards collaborative robotics in top view surveillance: A framework for multiple object tracking by detection using deep learning

    Imran Ahmed, Sadia Din, Gwanggil Jeon, Francesco Pic- cialli, and Giancarlo Fortino. Towards collaborative robotics in top view surveillance: A framework for multiple object tracking by detection using deep learning. IEEE/CAA Jour- nal of Automatica Sinica, 8(7):1253–1270, 2021. 1

    work page 2021

  3. [3]

    Bakhtiar Hasan, A

    Mk Bashar, Samia Islam, Kashifa Kawaakib Hussain, Md. Bakhtiar Hasan, A. B. M. Ashikur Rahman, and Md. Hasanul Kabir. Multiple object tracking in recent times: A literature review, 2022. 1 8

    work page 2022

  4. [4]

    Evaluating mul- tiple object tracking performance: The clear mot metrics

    Keni Bernardin and Rainer Stiefelhagen. Evaluating mul- tiple object tracking performance: The clear mot metrics. EURASIP Journal on Image and Video Processing, 2008, 01

    work page 2008

  5. [5]

    Simple online and realtime tracking

    Alex Bewley, Zongyuan Ge, Lionel Ott, Fabio Ramos, and Ben Upcroft. Simple online and realtime tracking. In 2016 IEEE International Conference on Image Processing (ICIP). IEEE, Sept. 2016. 2

    work page 2016

  6. [6]

    Yolov4: Optimal speed and accuracy of object detection, 2020

    Alexey Bochkovskiy, Chien-Yao Wang, and Hong- Yuan Mark Liao. Yolov4: Optimal speed and accuracy of object detection, 2020. 2, 4

    work page 2020

  7. [7]

    A unifying mutual information view of metric learn- ing: cross-entropy vs

    Malik Boudiaf, J ´erˆome Rony, Imtiaz Masud Ziko, Eric Granger, Marco Pedersoli, Pablo Piantanida, and Ismail Ben Ayed. A unifying mutual information view of metric learn- ing: cross-entropy vs. pairwise losses, 2021. 4

    work page 2021

  8. [8]

    Ross Beveridge, and Stephen O’Hara

    Mohamed Chaabane, Peter Zhang, J. Ross Beveridge, and Stephen O’Hara. Deft: Detection embeddings for tracking,

    work page

Show all 71 references
  1. [9]

    Large scale online learning of image similarity through rank- ing

    Gal Chechik, Varun Sharma, Uri Shalit, and Samy Bengio. Large scale online learning of image similarity through rank- ing. Journal of Machine Learning Research, 11(Mar):1109– 1135, 2010. 2

    2010

  2. [10]

    A simple framework for contrastive learning of visual representations, 2020

    Ting Chen, Simon Kornblith, Mohammad Norouzi, and Ge- offrey Hinton. A simple framework for contrastive learning of visual representations, 2020. 4

    2020

  3. [11]

    Deep learning in video multi-object tracking: A survey.Neu- rocomputing, 381:61–88, Mar

    Gioele Ciaparrone, Francisco Luque S ´anchez, Siham Tabik, Luigi Troiano, Roberto Tagliaferri, and Francisco Herrera. Deep learning in video multi-object tracking: A survey.Neu- rocomputing, 381:61–88, Mar. 2020. 1

    2020

  4. [12]

    Soccernet-tracking: Multiple object track- ing dataset and benchmark in soccer videos, 2022

    Anthony Cioppa, Silvio Giancola, Adrien Deliege, Le Kang, Xin Zhou, Zhiyu Cheng, Bernard Ghanem, and Marc Van Droogenbroeck. Soccernet-tracking: Multiple object track- ing dataset and benchmark in soccer videos, 2022. 1

    2022

  5. [13]

    Motchallenge: A benchmark for single-camera multiple target tracking, 2020

    Patrick Dendorfer, Aljo ˇsa O ˇsep, Anton Milan, Konrad Schindler, Daniel Cremers, Ian Reid, Stefan Roth, and Laura Leal-Taix´e. Motchallenge: A benchmark for single-camera multiple target tracking, 2020. 5

    2020

  6. [14]

    Mot20: A benchmark for multi object tracking in crowded scenes, 2020

    Patrick Dendorfer, Hamid Rezatofighi, Anton Milan, Javen Shi, Daniel Cremers, Ian Reid, Stefan Roth, Konrad Schindler, and Laura Leal-Taix´e. Mot20: A benchmark for multi object tracking in crowded scenes, 2020. 5

    2020

  7. [15]

    Pedestrian detection: A benchmark

    Piotr Dollar, Christian Wojek, Bernt Schiele, and Pietro Per- ona. Pedestrian detection: A benchmark. In 2009 IEEE Con- ference on Computer Vision and Pattern Recognition, pages 304–311, 2009. 5

    2009

  8. [16]

    Strongsort: Make deep- sort great again, 2023

    Yunhao Du, Zhicheng Zhao, Yang Song, Yanyun Zhao, Fei Su, Tao Gong, and Hongying Meng. Strongsort: Make deep- sort great again, 2023. 5

    2023

  9. [17]

    A mobile vision system for robust multi-person tracking

    Andreas Ess, Bastian Leibe, Konrad Schindler, and Luc Van Gool. A mobile vision system for robust multi-person tracking. In 2008 IEEE Conference on Computer Vision and Pattern Recognition, pages 1–8, 2008. 5

    2008

  10. [18]

    Huang, Jiangmiao Pang, Linlu Qiu, Haofeng Chen, Trevor Darrell, and Fisher Yu

    Tobias Fischer, Thomas E. Huang, Jiangmiao Pang, Linlu Qiu, Haofeng Chen, Trevor Darrell, and Fisher Yu. Qdtrack: Quasi-dense similarity learning for appearance-only multiple object tracking, 2023. 1, 3, 8

    2023

  11. [19]

    Unsuper- vised pre-training for person re-identification, 2021

    Dengpan Fu, Dongdong Chen, Jianmin Bao, Hao Yang, Lu Yuan, Lei Zhang, Houqiang Li, and Dong Chen. Unsuper- vised pre-training for person re-identification, 2021. 4

    2021

  12. [20]

    Yolox: Exceeding yolo series in 2021, 2021

    Zheng Ge, Songtao Liu, Feng Wang, Zeming Li, and Jian Sun. Yolox: Exceeding yolo series in 2021, 2021. 2

    2021

  13. [21]

    Momentum contrast for unsupervised visual rep- resentation learning, 2020

    Kaiming He, Haoqi Fan, Yuxin Wu, Saining Xie, and Ross Girshick. Momentum contrast for unsupervised visual rep- resentation learning, 2020. 4

    2020

  14. [22]

    Fastreid: A pytorch toolbox for general instance re-identification

    Lingxiao He, Xingyu Liao, Wu Liu, Xinchen Liu, Peng Cheng, and Tao Mei. Fastreid: A pytorch toolbox for general instance re-identification. arXiv preprint arXiv:2006.02631,

    2006 arXiv

  15. [23]

    In de- fense of the triplet loss for person re-identification, 2017

    Alexander Hermans, Lucas Beyer, and Bastian Leibe. In de- fense of the triplet loss for person re-identification, 2017. 4

    2017

  16. [24]

    Joint monocular 3d vehicle detection and tracking, 2019

    Hou-Ning Hu, Qi-Zhi Cai, Dequan Wang, Ji Lin, Min Sun, Philipp Kr ¨ahenb¨uhl, Trevor Darrell, and Fisher Yu. Joint monocular 3d vehicle detection and tracking, 2019. 1

    2019

  17. [25]

    Ultralytics yolov5, 2020

    Glenn Jocher. Ultralytics yolov5, 2020. 3

    2020

  18. [26]

    Ultralytics yolo11, 2024

    Glenn Jocher and Jing Qiu. Ultralytics yolo11, 2024. 2

    2024

  19. [27]

    Ultralytics YOLO, Jan

    Glenn Jocher, Jing Qiu, and Ayush Chaurasia. Ultralytics YOLO, Jan. 2023. 2

    2023

  20. [28]

    Conftrack: Kalman filter-based multi- person tracking by utilizing confidence score of detection box

    Hyeonchul Jung, Seokjun Kang, Takgen Kim, and HyeongKi Kim. Conftrack: Kalman filter-based multi- person tracking by utilizing confidence score of detection box. In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, pages 6583–6592, 2024. 1, 3

    2024

  21. [29]

    Multi- ple object tracking for football game analysis

    Zoran Kalafati ´c, Tomislav Hrka ´c, and Karla Brki ´c. Multi- ple object tracking for football game analysis. In 2022 45th Jubilee International Convention on Information, Communi- cation and Electronic Technology (MIPRO), pages 936–941,

    2022

  22. [30]

    R. E. Kalman. A new approach to linear filtering and predic- tion problems. Journal of Basic Engineering , 82(1):35–45, 03 1960. 2

    1960

  23. [31]

    H. W. Kuhn. The hungarian method for the assignment problem. Naval Research Logistics Quarterly, 2(1-2):83–97,

  24. [32]

    Simpletrack: Re- thinking and improving the jde approach for multi-object tracking, 2022

    Jiaxin Li, Yan Ding, and Hualiang Wei. Simpletrack: Re- thinking and improving the jde approach for multi-object tracking, 2022. 1, 3, 8

    2022

  25. [33]

    One more check: Making ”fake background” be tracked again, 2021

    Chao Liang, Zhipeng Zhang, Xue Zhou, Bing Li, and Weim- ing Hu. One more check: Making ”fake background” be tracked again, 2021. 1, 3, 8

    2021

  26. [34]

    Rethinking the competition between detection and reid in multi-object tracking, 2022

    Chao Liang, Zhipeng Zhang, Xue Zhou, Bing Li, Shuyuan Zhu, and Weiming Hu. Rethinking the competition between detection and reid in multi-object tracking, 2022. 1, 3, 8

    2022

  27. [35]

    Focal loss for dense object detection, 2018

    Tsung-Yi Lin, Priya Goyal, Ross Girshick, Kaiming He, and Piotr Doll´ar. Focal loss for dense object detection, 2018. 3

    2018

  28. [36]

    Lawrence Zitnick, and Piotr Doll ´ar

    Tsung-Yi Lin, Michael Maire, Serge Belongie, Lubomir Bourdev, Ross Girshick, James Hays, Pietro Perona, Deva Ramanan, C. Lawrence Zitnick, and Piotr Doll ´ar. Microsoft coco: Common objects in context, 2015. 5

    2015

  29. [37]

    Unsupervised person re-identification via softened similarity learning

    Yutian Lin, Lingxi Xie, Yu Wu, Chenggang Yan, and Qi Tian. Unsupervised person re-identification via softened similarity learning. In 2020 IEEE/CVF Conference on Com- puter Vision and Pattern Recognition (CVPR), pages 3387– 3396, 2020. 4 9

    2020

  30. [38]

    Retinatrack: Online single stage joint detection and tracking, 2020

    Zhichao Lu, Vivek Rathod, Ronny V otel, and Jonathan Huang. Retinatrack: Online single stage joint detection and tracking, 2020. 1, 3

    2020

  31. [39]

    Hota: A higher order metric for evaluating multi- object tracking

    Jonathon Luiten, Aljo ˘sa O ˘sep, Patrick Dendorfer, Philip Torr, Andreas Geiger, Laura Leal-Taix ´e, and Bastian Leibe. Hota: A higher order metric for evaluating multi- object tracking. International Journal of Computer Vision , 129(2):548–578, Oct. 2020. 5

    2020

  32. [40]

    Mot16: A benchmark for multi-object tracking, 2016

    Anton Milan, Laura Leal-Taixe, Ian Reid, Stefan Roth, and Konrad Schindler. Mot16: A benchmark for multi-object tracking, 2016. 5

    2016

  33. [41]

    Belongie, and Ser-Nam Lim

    Kevin Musgrave, Serge J. Belongie, and Ser-Nam Lim. Py- torch metric learning. ArXiv, abs/2008.09164, 2020. 5

    2008 arXiv

  34. [42]

    Sort and deep-sort based multi-object tracking for mobile robotics: Evaluation with new data association metrics

    Ricardo Pereira, Guilherme Carvalho, Lu ´ıs Garrote, and Ur- bano Nunes. Sort and deep-sort based multi-object tracking for mobile robotics: Evaluation with new data association metrics. Applied Sciences, 12:1319, 01 2022. 1

    2022

  35. [43]

    Yolov3: An incremental improvement, 2018

    Joseph Redmon and Ali Farhadi. Yolov3: An incremental improvement, 2018. 3

    2018

  36. [44]

    Faster r-cnn: Towards real-time object detection with region proposal networks, 2016

    Shaoqing Ren, Kaiming He, Ross Girshick, and Jian Sun. Faster r-cnn: Towards real-time object detection with region proposal networks, 2016. 2

    2016

  37. [45]

    Joint counting, detection and re- identification for multi-object tracking, 2024

    Weihong Ren, Denglu Wu, Hui Cao, Xi’ai Chen, Zhi Han, and Honghai Liu. Joint counting, detection and re- identification for multi-object tracking, 2024. 1, 3, 8

    2024

  38. [46]

    Zou, Rita Cuc- chiara, and Carlo Tomasi

    Ergys Ristani, Francesco Solera, Roger S. Zou, Rita Cuc- chiara, and Carlo Tomasi. Performance measures and a data set for multi-target, multi-camera tracking, 2016. 5

    2016

  39. [47]

    Rousseeuw

    Peter J. Rousseeuw. Silhouettes: A graphical aid to the inter- pretation and validation of cluster analysis. Journal of Com- putational and Applied Mathematics, 20:53–65, 1987. 5

    1987

  40. [48]

    Facenet: A unified embedding for face recognition and clus- tering

    Florian Schroff, Dmitry Kalenichenko, and James Philbin. Facenet: A unified embedding for face recognition and clus- tering. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 815–823, 2015. 2

    2015

  41. [49]

    Crowdhuman: A benchmark for detecting human in a crowd, 2018

    Shuai Shao, Zijian Zhao, Boxun Li, Tete Xiao, Gang Yu, Xiangyu Zhang, and Jian Sun. Crowdhuman: A benchmark for detecting human in a crowd, 2018. 5

    2018

  42. [50]

    Boost- track: boosting the similarity measure and detection confi- dence for improved multiple object tracking.Machine Vision and Applications, 35(3), 2024

    Vukasin D Stanojevic and Branimir T Todorovic. Boost- track: boosting the similarity measure and detection confi- dence for improved multiple object tracking.Machine Vision and Applications, 35(3), 2024. 1, 3, 5

    2024

  43. [51]

    Online and real-time tracking in a surveillance scenario, 2021

    Oliver Urbann, Oliver Bredtmann, Maximilian Otten, Jan- Philip Richter, Thilo Bauer, and David Zibriczky. Online and real-time tracking in a surveillance scenario, 2021. 1

    2021

  44. [52]

    Repre- sentation learning with contrastive predictive coding, 2019

    Aaron van den Oord, Yazhe Li, and Oriol Vinyals. Repre- sentation learning with contrastive predictive coding, 2019. 3, 4

    2019

  45. [53]

    Clausi, and John Zelek

    Kanav Vats, Pascale Walters, Mehrnaz Fani, David A. Clausi, and John Zelek. Player tracking and identification in ice hockey, 2021. 1

    2021

  46. [54]

    Deep metric learning with angular loss, 2017

    Jian Wang, Feng Zhou, Shilei Wen, Xiao Liu, and Yuanqing Lin. Deep metric learning with angular loss, 2017. 4

    2017

  47. [55]

    Xun Wang, Xintong Han, Weilin Huang, Dengke Dong, and Matthew R. Scott. Multi-similarity loss with general pair weighting for deep metric learning, 2020. 4

    2020

  48. [56]

    Smiletrack: Simi- larity learning for occlusion-aware multiple object tracking,

    Yu-Hsiang Wang, Jun-Wei Hsieh, Ping-Yang Chen, Ming- Ching Chang, Hung Hin So, and Xin Li. Smiletrack: Simi- larity learning for occlusion-aware multiple object tracking,

  49. [57]

    Towards real-time multi-object tracking,

    Zhongdao Wang, Liang Zheng, Yixuan Liu, Yali Li, and Shengjin Wang. Towards real-time multi-object tracking,

  50. [58]

    Justin Wilson and Ming C. Lin. Avot: Audio-visual object tracking of multiple objects for robotics. In 2020 IEEE In- ternational Conference on Robotics and Automation (ICRA), pages 10045–10051, 2020. 1

    2020

  51. [59]

    Simple online and realtime tracking with a deep association metric,

    Nicolai Wojke, Alex Bewley, and Dietrich Paulus. Simple online and realtime tracking with a deep association metric,

  52. [60]

    Joint detection and identification feature learn- ing for person search, 2017

    Tong Xiao, Shuang Li, Bochao Wang, Liang Lin, and Xiao- gang Wang. Joint detection and identification feature learn- ing for person search, 2017. 5

    2017

  53. [61]

    Hard to track objects with irregular motions and sim- ilar appearances? make it easier by buffering the matching space, 2023

    Fan Yang, Shigeyuki Odashima, Shoichi Masui, and Shan Jiang. Hard to track objects with irregular motions and sim- ilar appearances? make it easier by buffering the matching space, 2023. 3

    2023

  54. [62]

    Utm: A unified multiple object tracking model with identity- aware feature enhancement

    Sisi You, Hantao Yao, Bing-kun Bao, and Changsheng Xu. Utm: A unified multiple object tracking model with identity- aware feature enhancement. In 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 21876–21886, 2023. 3

    2023

  55. [63]

    Relationtrack: Relation-aware multiple object tracking with decoupled representation, 2021

    En Yu, Zhuoling Li, Shoudong Han, and Hongwei Wang. Relationtrack: Relation-aware multiple object tracking with decoupled representation, 2021. 1, 3, 8

    2021

  56. [64]

    Poi: Multiple object tracking with high per- formance detection and appearance feature, 2016

    Fengwei Yu, Wenbo Li, Quanquan Li, Yu Liu, Xiaohua Shi, and Junjie Yan. Poi: Multiple object tracking with high per- formance detection and appearance feature, 2016. 3

    2016

  57. [65]

    Dauphin, and David Lopez-Paz

    Hongyi Zhang, Moustapha Cisse, Yann N. Dauphin, and David Lopez-Paz. mixup: Beyond empirical risk minimiza- tion, 2018. 3

    2018

  58. [66]

    Citypersons: A diverse dataset for pedestrian detection,

    Shanshan Zhang, Rodrigo Benenson, and Bernt Schiele. Citypersons: A diverse dataset for pedestrian detection,

  59. [67]

    Tempo- ral correlation meets embedding: Towards a 2nd generation of jde-based real-time multi-object tracking, 2024

    Yunfei Zhang, Chao Liang, Jin Gao, Zhipeng Zhang, Weim- ing Hu, Stephen Maybank, Xue Zhou, and Liang Li. Tempo- ral correlation meets embedding: Towards a 2nd generation of jde-based real-time multi-object tracking, 2024. 1, 3, 8

    2024

  60. [68]

    Bytetrack: Multi-object tracking by associating every detection box, 2022

    Yifu Zhang, Peize Sun, Yi Jiang, Dongdong Yu, Fucheng Weng, Zehuan Yuan, Ping Luo, Wenyu Liu, and Xinggang Wang. Bytetrack: Multi-object tracking by associating every detection box, 2022. 2

    2022

  61. [69]

    Fairmot: On the fairness of detection and re- identification in multiple object tracking

    Yifu Zhang, Chunyu Wang, Xinggang Wang, Wenjun Zeng, and Wenyu Liu. Fairmot: On the fairness of detection and re- identification in multiple object tracking. International Jour- nal of Computer Vision, 129(11):3069–3087, Sept. 2021. 1, 3, 8

    2021

  62. [70]

    Person re-identification in the wild, 2017

    Liang Zheng, Hengheng Zhang, Shaoyan Sun, Manmohan Chandraker, Yi Yang, and Qi Tian. Person re-identification in the wild, 2017. 5

    2017

  63. [71]

    Ob- jects as points, 2019

    Xingyi Zhou, Dequan Wang, and Philipp Kr ¨ahenb¨uhl. Ob- jects as points, 2019. 2 10

    2019

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.