Pith. sign in

REVIEW 4 major objections 9 minor 73 references

OCDet: Object Center Detection via Bounding Box-Aware Heatmap Prediction on Edge Devices with NPUs

T0 review · 4 major / 9 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read Bounding-box-aware center heatmaps beat YOLO11 for coarse object localization on edge NPUs, this paper argues.

desk verdict Genuinely practical edge-object-centering work with real NPU latency data, but the headline CAS numbers rest on a metric that is not fully specified as printed. read the letter →

arxiv 2411.15653 v1 pith:KD73YMEO submitted 2024-11-23 cs.CV

Chen Xin , Thomas Motz , Andreas Hartel , Enkelejda Kasneci This is my paper · ORCID
classification cs.CV
keywords objectcenterdetectionheatmappredictionedgeNPUGeneralizedCenternessBalancedContinuousFocalLossAlignmentScoreMobileNetV4SemanticFPN
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

OCDet claims that for coarse object localization on edge devices with NPUs, predicting object center heatmaps is more effective than full bounding-box detection or keypoint detection. The framework combines Generalized Centerness (GC), a bounding-box-derived probability map, with Balanced Continuous Focal Loss (BCFL) and a lightweight Semantic FPN with MobileNetV4 backbones. Under the paper's Center Alignment Score (CAS) metric, OCDet is reported to exceed YOLO11 by up to 23% while using 42% fewer parameters, 34% less computation, and 64% lower NPU latency, and to beat a keypoint-detection baseline by up to 186% with identical models. If these numbers hold, surveillance, industrial automation, and other high-recall edge applications could get more reliable coarse localization at lower hardware cost.

What carries the argument

The load-bearing object is Generalized Centerness (GC), the identity GC(x,y) = (min(l,r)/max(l,r))^eta times (min(t,b)/max(t,b))^phi, which converts a bounding box into a [0,1] per-pixel probability heatmap whose concentration is tuned by eta and $\varphi$ (eta=phi=0.5 reproduces the Centerness of FCOS). It replaces the fixed-Gaussian ground truth of keypoint frameworks. Two supporting mechanisms carry the training and evaluation: Balanced Continuous Focal Loss (BCFL), which reweights Quality Focal Loss by an adaptive factor alpha_c(y) = $\alpha$*y + (1-$\alpha$)*(1-y) to suppress low-value background points and emphasize high-value center regions; and the Center Alignment Score (CAS), which applies Hungarian matching and subtracts a Matched Discrepancy and a Cardinality Penalty from 1. The models are built on Semantic FPN with MobileNetV4, chosen for NPU-friendly operations.

What would settle it

Run the released code on the COCO val2017 set, compute Eq. (9) with N_i interpreted as the number of images, and check whether the resulting per-image penalties (CP around 0.5-0.65 for OCDet variants in Tab. 1) are consistent with typical object counts per image (roughly 7-10); if the metric instead normalizes by object count or matched pairs, the tabulated CAS values change and the headline percentages against YOLO11 need recomputation. A direct way to settle it: reproduce a single row of Tab. 1 from the public code and compare the computed CAS to the printed 0.297.

Watch

Extended reading notes

Core claim

On the paper's own terms, the discovery is that the standard practice of placing a fixed Gaussian at each target point, and the standard practice of regressing bounding boxes, are both mismatched to the object-center detection task. OCDet replaces them with Generalized Centerness, which assigns every pixel inside a box a probability of being the center based on the ratio of distances to the box edges, with shape controlled by hyperparameters eta and phi. Under this supervision, models trained with the proposed Balanced Continuous Focal Loss learn heatmaps that better preserve central geometry, and the paper reports that this combination achieves consistently higher Center Alignment Scores than YOLO11 and YOLOv8 across all five model variants, with substantially smaller models and lower latency, and higher recall and F1.

Load-bearing premise

The entire comparison rests on the Center Alignment Score being a well-defined, identically normalized metric across all models; Eq. (9) of the paper sums CP_i/N_i and MD_i/N_i without ever defining the count N_i, so the reported CAS values are only meaningful if that normalization is fixed in a way the paper does not state.

Editorial extensions

If this is right

  • If CAS is the right yardstick, then object center detection on COCO-like data is better served by heatmap regression with GC ground truth than by YOLO-family box regression, at least under the reported protocol: up to 23% higher CAS with 42% fewer parameters and 64% lower NPU latency.
  • Replacing the fixed Gaussian ground truth of keypoint frameworks with GC is predicted to raise CAS by up to 186% at identical model size, with the largest relative gains at the smallest model scale (OCDet-N versus KP-N).
  • The ablation results imply that BCFL is the component that unlocks small-object center detection: CASS jumps from 0.044 (QFL) to 0.314 (BCFL) in the loss comparison table.
  • Because GC only needs bounding boxes, the framework transfers to any object detection dataset and to automated box-labeling pipelines without extra manual keypoint annotation.
  • A continuous, scale-aware metric like CAS could complement or replace threshold-based AP for center-focused localization tasks, since it avoids the abrupt performance changes induced by fixed IoU thresholds.

Reading between the lines

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

  • A testable extension would be to attach GC ground truth to an existing CenterNet-style detector and compare against the fixed-Gaussian variant, isolating whether the reported gains come from the heatmap shape or from the FPN+BCFL combination.
  • A scale-free version of CAS could divide the cardinality penalty by the per-image ground-truth count; such normalization would remove the N_i ambiguity and make cross-dataset comparisons safer.
  • Learning eta and phi per class is a natural next step: the ablation's finding that skis prefer eta=0.9/phi=0.1 and bottles prefer eta=0.1/phi=0.9 suggests a per-class shape model could recover the small-object CASS deficit.
  • On the hardware side, the paper's exclusion of SE blocks and LayerNorm because of NPU latency is consistent with a simple latency-overhead intuition; a direct per-operation timing measurement on the i.MX 8M Plus would confirm how much of the 64% latency win is architectural versus quantization-related.
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 / 9 minor

Summary. The paper introduces OCDet, a heatmap-based object center detection framework for NPU-equipped edge devices. Ground-truth heatmaps are generated from bounding-box annotations via a Generalized Centerness (GC) map with shape parameters eta and phi; the models use MobileNetV4 backbones with Semantic FPN and are trained with a Balanced Continuous Focal Loss (BCFL). Evaluation is performed with a proposed Center Alignment Score (CAS) that uses Hungarian matching, a matched-discrepancy term, and a cardinality penalty. On COCO, the authors report that OCDet variants achieve up to 23% higher CAS than YOLO11 while using 42% fewer parameters, 34% less computation, and 64% lower NPU latency, and up to 186% higher CAS than identical models trained under a standard keypoint-detection framework. The paper includes ablations over ground-truth type, GC exponents, loss variants, BCFL hyperparameters, backbone, architecture, output stride, FPN modules, and peak-identification settings.

Significance. If the quantitative claims survive scrutiny, this is a practically useful contribution: it defines a cheap edge-deployable center-detection task, proposes a box-aware heatmap representation, and provides extensive ablations, NPU latency measurements, and released code. The strengths are real: the architecture and loss design are reasonable, the ablations are informative, and the qualitative failure cases are honestly discussed. However, the central quantitative claims are currently not verifiable because the CAS metric is incompletely specified, and the evaluation protocol has selection-on-validation issues. The significance of the specific numerical advantages over YOLO11 and the keypoint baselines is contingent on fixing the metric definition and adding held-out evaluation with variance estimates.

major comments (4)
  1. [Sec. 3.5, Eqs. (7)-(9)] Eq. (7) defines MD_i as the 'average' of normalized matched distances but writes only a sum, omitting the division by |M|. Eq. (9) then subtracts sum CP_i/N_i and sum MD_i/N_i without ever defining N_i. If N_i is the number of images, the CP term is a mean raw unmatched-point count, which is neither normalized nor bounded and would make the reported CAS values in Tab. 1 (0.2-0.4) impossible for typical COCO image object counts; if N_i is instead an object count or a per-image normalization, that definition is absent. The text also provides no pseudocode or code excerpt for the metric, so the GitHub link alone does not pin down the formula used to produce Table 1. Because every headline comparison—including the 23% CAS improvement over YOLO11 and the 186% over the KP baselines—is expressed in CAS units, the central claim cannot be checked as written. Please specify N_i, the averaging in MD_i (including the case |M|=0), and the exact per-image normalization, and verify that the released code reproduces Table 1.
  2. [Sec. 4.1, Sec. 4.3, and Appendix E] The validation set val2017 is used both to select hyperparameters and to report final results. Specifically, Tab. 5 selects the probability threshold (0.6) and alpha for BCFL on the person center-detection task using CAS, Tab. 6 selects gamma, and Appendix E selects the peak-identification probability threshold (0.5) and minimum distance (0.3) on OCDet-M using CAS; these choices are then applied across all models in Tab. 1. This creates a selection loop: the same data and metric that define the reported superiority are used to tune the method. At minimum, report results on a held-out split or use nested cross-validation, and quantify the selection bias. Without this, the 'up to 23%' claim may be an optimistic estimate.
  3. [Table 1 and Sec. 4.2] The comparison with YOLOv8 and YOLO11 extracts centers from bounding-box outputs of off-the-shelf detectors, but no confidence threshold, NMS setting, or other center-extraction protocol is reported for these baselines, and the detectors are not trained or adapted for the center-detection objective. YOLO models are optimized for box IoU, not center alignment, so the comparison conflates task definition with model capability. Since the claimed Pareto dominance in Fig. 1 depends on this protocol, please provide a sensitivity analysis over YOLO confidence thresholds and report the exact thresholds used, or explicitly reframe the claims as 'center extraction from off-the-shelf detectors' rather than 'OCDet outperforms YOLO11 in center detection.'
  4. [Tables 1-10] No error bars or multiple-run statistics are reported. Several headline differences are small in absolute terms (e.g., OCDet-S 0.313 vs YOLO11s 0.309; OCDet-M 0.362 vs YOLOv8m 0.357), and without seed variance it is impossible to determine whether the 1% and 6% improvements are significant. Please report mean +/- std over at least three seeds for the main comparison table and for the key ablations (Tabs. 1, 2, 4, 5, 6), and state the number of runs used.
minor comments (9)
  1. [Sec. 3.5, Eq. (7)] MD_i is called an average but the division by |M| is missing; please add the denominator and handle the case |M|=0 explicitly.
  2. [Sec. 3.5, Eq. (9)] The summation indices and the definition of N_i are ambiguous; use consistent notation, for example CAS = 1 - (1/I) * sum_i (CP_i + MD_i), and define all quantities.
  3. [Sec. 3.5, text] The phrase 'Matched Distance (MD) and Coverage Precision (CP)' should be 'Matched Discrepancy and Cardinality Penalty' to match Eq. (8) and the rest of the paper.
  4. [Sec. 3.4, Eq. (5)] The permutation notation sigma in Sigma_N with the assumption N > G is only valid when predictions outnumber ground truths; if G > N, the matching should be defined over injections. Please state the general case.
  5. [Sec. 4.3 and Appendix E] The ablation tables (Tabs. 2-6 and 7-10) are for person center detection only, but the selected hyperparameters are then applied to the 80-class COCO task; please justify the transferability or provide 80-class ablations.
  6. [Sec. 4.1 and Appendix D] The '2x schedule' with 24 epochs is nonstandard; please clarify the schedule definition.
  7. [References] Reference [56] for FCOS appears to point to a paper on 3D object detection on LiDAR range images; the correct FCOS reference (Tian et al., ICCV 2019) should be cited.
  8. [Sec. 1 and Sec. 4.2] There are typos: 'CAS grains' should be 'CAS gains', and the final sentence of Sec. 4.2 contains a comma splice; please proofread.
  9. [Sec. 4.1 and Appendix E] The 'probability threshold' of 0.6 used to derive alpha and the peak-identification probability threshold of 0.5 are different concepts but are both called 'probability threshold'; please use distinct terms to avoid confusion.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: OCDet's claimed gains are grounded in an external COCO benchmark and standard baselines, with the proposed CAS metric serving as an evaluation measure rather than as an input to the method.

full rationale

The paper's derivation chain is self-contained and externally grounded. Generalized Centerness (GC) is defined directly from bounding-box geometry (Eq. 1), Balanced Continuous Focal Loss (BCFL) is an explicit modification of Quality Focal Loss (Eqs. 3-4), and the OCDet models are built on MobileNetV4 backbones and Semantic FPN. The Center Alignment Score (CAS) is a proposed evaluation metric computed by Hungarian matching (Eqs. 5-9); it is not used as a training loss or as a constraint in model construction, so the headline comparisons against YOLO11 and keypoint baselines are not forced by definition. Hyperparameter choices (eta, phi, alpha, gamma, peak thresholds) are selected through CAS ablations, but this is conventional model selection on a validation protocol, not a fitted parameter renamed as a prediction: the reported numbers still come from independent forward inference of the trained models on COCO. The only self-citation, reference [66] for automated bounding box labeling, is motivational and non-load-bearing. The undefined N_i in Eq. (9) and the missing normalization in Eq. (7) are genuine reproducibility and correctness concerns that should be addressed by the authors, but they do not make the derivation itself circular: they make the metric ambiguous, not equivalent to its inputs. No step in the paper reduces, by its own equations, to a prior fit or to a self-citation chain.

Assumptions & free parameters 6 free parameters · 5 assumptions · 0 invented entities

The framework introduces no new physical entities. The free parameters are hyperparameters or evaluation choices tuned on validation data; the axioms are standard math or domain assumptions required for the reported empirical results. The invented_entities list is empty because CAS, BCFL, and GC are methods/representations, not newly postulated physical things.

free parameters (6)
  • GC exponents eta, phi = 0.5, 0.5 (default)
    Chosen default, ablation Tab. 3 shows 0.7/0.7 comparable; directly controls GT heatmap shape and influences all results.
  • BCFL gamma = 2
    Selected via ablation Tab. 6; balances hard negative emphasis.
  • BCFL alpha = 0.984 (80-class), 0.964 (person)
    Derived from inverse class frequency with probability threshold 0.6; threshold tuned on validation set (Tab. 5).
  • Peak identification probability threshold = 0.5
    Tuned on validation set (Fig. 7a).
  • Peak identification minimum distance = 0.3
    Tuned on validation set (Fig. 7b).
  • Hungarian matching weights lambda, mu = 1, 1
    Set manually in Sec. 3.4; not ablated.
assumptions (5)
  • domain assumption Object center is defined as the center of its bounding box.
    Stated in Sec. 1 and used throughout; applications needing coarse localization may prefer this definition, but it is not universally true for rotated/occluded objects.
  • ad hoc to paper The CAS metric with Hungarian matching and normalized distance is a valid measure of center detection quality.
    Defined in Sec. 3.5; used as the primary evaluation, with ambiguous normalization in Eq. (9).
  • domain assumption GC heatmaps encoding per-pixel probability of being a center are an appropriate training target.
    Introduced in Sec. 3.1; relies on FCOS centerness notion extended to heatmaps.
  • domain assumption NPU latency measured on i.MX 8M Plus with uint8 TFLite is representative of edge NPU performance.
    Used for all latency comparisons in Tab. 1; no other hardware considered.
  • standard math Standard focal loss and quality focal loss formulations as cited provide the correct functional form.
    BCFL builds directly on QFL (Eq. 3) and FL (Eq. 2) from [49,33].

how reviews work

0 comments
Cite this review

Pith. "Pith review of OCDet: Object Center Detection via Bounding Box-Aware Heatmap Prediction on Edge Devices with NPUs." pith.science (2026). https://pith.science/paper/KD73YMEO

@misc{pith2026241115653,
  author       = {Pith},
  title        = {Pith review of: OCDet: Object Center Detection via Bounding Box-Aware Heatmap Prediction on Edge Devices with NPUs},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/KD73YMEO}},
  note         = {Machine review of arXiv:2411.15653}
}
read the original abstract

Real-time object localization on edge devices is fundamental for numerous applications, ranging from surveillance to industrial automation. Traditional frameworks, such as object detection, segmentation, and keypoint detection, struggle in resource-constrained environments, often resulting in substantial target omissions. To address these challenges, we introduce OCDet, a lightweight Object Center Detection framework optimized for edge devices with NPUs. OCDet predicts heatmaps representing object center probabilities and extracts center points through peak identification. Unlike prior methods using fixed Gaussian distribution, we introduce Generalized Centerness (GC) to generate ground truth heatmaps from bounding box annotations, providing finer spatial details without additional manual labeling. Built on NPU-friendly Semantic FPN with MobileNetV4 backbones, OCDet models are trained by our Balanced Continuous Focal Loss (BCFL), which alleviates data imbalance and focuses training on hard negative examples for probability regression tasks. Leveraging the novel Center Alignment Score (CAS) with Hungarian matching, we demonstrate that OCDet consistently outperforms YOLO11 in object center detection, achieving up to 23% higher CAS while requiring 42% fewer parameters, 34% less computation, and 64% lower NPU latency. When compared to keypoint detection frameworks, OCDet achieves substantial CAS improvements up to 186% using identical models. By integrating GC, BCFL, and CAS, OCDet establishes a new paradigm for efficient and robust object center detection on edge devices with NPUs. The code is released at https://github.com/chen-xin-94/ocdet.

Figures

Figures reproduced from arXiv: 2411.15653 by the authors.

Figure 1
Figure 1. Performance comparison between our proposed OCDet [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Illustration of our proposed OCDet, an Object Center Detection framework, comprising: [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. (b), where BCFL considerably reduces the loss for lower target values while preserving relatively larger losses for high-value targets, thereby accentuating points near the object centers. 0.0 0.2 0.4 0.6 0.8 1.0 Target y 0.0 0.2 0.4 0.6 0.8 1.0 c = 0.5 = 0.75 = 0.9 = 1 (a) αc(y) under various α values 0.0 0.2 0.4 0.6 0.8 1.0 Target y 0.00 0.05 0.10 0.15 Loss QFL BCFL (b) BCFL vs QFL [PITH_FULL_IMAGE:figures/full_f… view at source ↗
Figures from the paper (7 more)
Figure 4
Figure 4. Figure 4: Illustration of Generalized Centerness heatmaps of a [PITH_FULL_IMAGE:figures/full_fig_p012_4.png]
Figure 5
Figure 5. Figure 5: Evaluation of the proposed OCDet against state-of-the-art real-time object detectors YOLOv8, YOLO11, and the same models [PITH_FULL_IMAGE:figures/full_fig_p013_5.png]
Figure 6
Figure 6. Figure 6: Architecture diagrams of (a) Semantic FPN [ [PITH_FULL_IMAGE:figures/full_fig_p014_6.png]
Figure 7
Figure 7. Figure 7: Visualization of the ablation study for peak identification [PITH_FULL_IMAGE:figures/full_fig_p016_7.png]
Figure 8
Figure 8. Figure 8: Comparison of OCDet-X (right), YOLO11x (center), and Ground Truth (left) for object center detection. Detected object centers [PITH_FULL_IMAGE:figures/full_fig_p017_8.png]
Figure 9
Figure 9. Figure 9: Failure cases of OCDet-X (right) in object center detection compared with Ground Truth (left) and YOLO11x (center). The first [PITH_FULL_IMAGE:figures/full_fig_p018_9.png]
Figure 10
Figure 10. Figure 10: More visualizations of OCDet-X’s object center detection results. [PITH_FULL_IMAGE:figures/full_fig_p019_10.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

73 extracted references · 54 canonical work pages

  1. [1]

    Hin- ton

    Jimmy Lei Ba, Jamie Ryan Kiros, and Geoffrey E. Hin- ton. Layer normalization. arXiv preprint arXiv:1607.06450,

    arXiv

  2. [2]

    Yolov4: Optimal speed and accuracy of object detection

    Alexey Bochkovskiy, Chien-Yao Wang, and Hong- Yuan Mark Liao. Yolov4: Optimal speed and accuracy of object detection. arXiv preprint arXiv:2004.10934, 2020. 2

    arXiv 2004

  3. [3]

    Realtime multi-person 2d pose estimation using part affinity fields

    Zhe Cao, Tomas Simon, Shih-En Wei, and Yaser Sheikh. Realtime multi-person 2d pose estimation using part affinity fields. In CVPR, pages 7291–7299, 2017. 1, 2

    work page 2017

  4. [4]

    End- to-end object detection with transformers

    Nicolas Carion, Francisco Massa, Gabriel Synnaeve, Nicolas Usunier, Alexander Kirillov, and Sergey Zagoruyko. End- to-end object detection with transformers. In ECCV, pages 213–229. Springer, 2020. 1

    2020

  5. [5]

    Human pose estimation with iterative error feedback

    Joao Carreira, Pulkit Agrawal, Katerina Fragkiadaki, and Ji- tendra Malik. Human pose estimation with iterative error feedback. In CVPR, pages 4733–4742, 2016. 3

    work page 2016

  6. [6]

    Deeplab: Semantic image segmentation with deep convolutional nets, atrous convolu- tion, and fully connected crfs

    Liang-Chieh Chen, George Papandreou, Iasonas Kokkinos, Kevin Murphy, and Alan L Yuille. Deeplab: Semantic image segmentation with deep convolutional nets, atrous convolu- tion, and fully connected crfs. IEEE TPAMI, 40(4):834–848,

    work page

  7. [7]

    Cascaded pyramid network for multi-person pose estimation

    Yilun Chen, Zhicheng Wang, Yuxiang Peng, Zhiqiang Zhang, Gang Yu, and Jian Sun. Cascaded pyramid network for multi-person pose estimation. In CVPR, pages 7103– 7112, 2018. 1, 2, 4

    work page 2018

  8. [8]

    Higherhrnet: Scale-aware representation learning for bottom-up human pose estima- tion

    Bowen Cheng, Bin Xiao, Jingdong Wang, Honghui Shi, Thomas S Huang, and Lei Zhang. Higherhrnet: Scale-aware representation learning for bottom-up human pose estima- tion. In CVPR, pages 5386–5395, 2020. 1, 2

    work page 2020

Show all 73 references
  1. [9]

    Masked-attention mask transformer for universal image segmentation

    Bowen Cheng, Ishan Misra, Alexander G Schwing, Alexan- der Kirillov, and Rohit Girdhar. Masked-attention mask transformer for universal image segmentation. In CVPR, pages 1290–1299, 2022. 2, 3

    2022

  2. [10]

    Centroidnet: A deep neural network for joint object localization and counting

    Klaas Dijkstra, Jaap van de Loosdrecht, Lambert RB Schomaker, and Marco A Wiering. Centroidnet: A deep neural network for joint object localization and counting. In Machine Learning and Knowledge Discovery in Databases: European Conference, ECML PKDD 2018, Dublin, Ireland, Sep...

    2018

  3. [11]

    Repvgg: Making vgg-style convnets great again

    Xiaohan Ding, Xiangyu Zhang, Ningning Ma, Jungong Han, Guiguang Ding, and Jian Sun. Repvgg: Making vgg-style convnets great again. In CVPR, pages 13733–13742, 2021. 3

    2021

  4. [12]

    Centroid based person detection using pixel- wise prediction of the position

    Petr Dolezel, Pavel Skrabanek, Dominik Stursa, Bruno Baruque Zanon, Hector Cogollos Adrian, and Pavel Kryda. Centroid based person detection using pixel- wise prediction of the position. Journal of Computational Science, 63:101760, 2022. 8

    2022

  5. [13]

    Centernet: Keypoint triplets for object detection

    Kaiwen Duan, Song Bai, Lingxi Xie, Honggang Qi, Qing- ming Huang, and Qi Tian. Centernet: Keypoint triplets for object detection. In ICCV, pages 6569–6578, 2019. 1, 2

    2019

  6. [14]

    Deep learning-based anomaly detection in video surveil- lance: A survey

    Huu-Thanh Duong, Viet-Tuan Le, and Vinh Truong Hoang. Deep learning-based anomaly detection in video surveil- lance: A survey. Sensors, 23(11):5024, 2023. 1

    2023

  7. [15]

    The pascal visual object classes challenge: A retrospective.IJCV, 111:98–136, 2015

    Mark Everingham, SM Ali Eslami, Luc Van Gool, Christo- pher KI Williams, John Winn, and Andrew Zisserman. The pascal visual object classes challenge: A retrospective.IJCV, 111:98–136, 2015. 2

    2015

  8. [16]

    Instagen: Enhancing object detection by training on synthetic dataset

    Chengjian Feng, Yujie Zhong, Zequn Jie, Weidi Xie, and Lin Ma. Instagen: Enhancing object detection by training on synthetic dataset. In CVPR, pages 14121–14130, 2024. 2

    2024

  9. [17]

    A review and comparative study on probabilistic ob- ject detection in autonomous driving

    Di Feng, Ali Harakeh, Steven L Waslander, and Klaus Diet- mayer. A review and comparative study on probabilistic ob- ject detection in autonomous driving. IEEE Transactions on Intelligent Transportation Systems, 23(8):9961–9980, 2021. 1

    2021

  10. [18]

    Bottom-up human pose estimation via disen- tangled keypoint regression

    Zigang Geng, Ke Sun, Bin Xiao, Zhaoxiang Zhang, and Jing- dong Wang. Bottom-up human pose estimation via disen- tangled keypoint regression. In CVPR, pages 14676–14686,

  11. [19]

    Delving deep into rectifiers: Surpassing human-level perfor- mance on imagenet classification

    Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. Delving deep into rectifiers: Surpassing human-level perfor- mance on imagenet classification. In CVPR, pages 1026– 1034, 2015. 2

    2015

  12. [20]

    Mask r-cnn

    Kaiming He, Georgia Gkioxari, Piotr Doll ´ar, and Ross Gir- shick. Mask r-cnn. In ICCV, pages 2961–2969, 2017. 2, 3

    2017

  13. [21]

    Gaussian error linear units (gelus)

    Dan Hendrycks and Kevin Gimpel. Gaussian error linear units (gelus). arXiv preprint arXiv:1606.08415, 2016. 4

    2016 arXiv

  14. [22]

    Searching for mo- bilenetv3

    Andrew Howard, Mark Sandler, Grace Chu, Liang-Chieh Chen, Bo Chen, Mingxing Tan, Weijun Wang, Yukun Zhu, Ruoming Pang, Vijay Vasudevan, et al. Searching for mo- bilenetv3. In CVPR, pages 1314–1324, 2019. 4, 3

    2019

  15. [23]

    Squeeze-and-excitation net- works

    Jie Hu, Li Shen, and Gang Sun. Squeeze-and-excitation net- works. In CVPR, pages 7132–7141, 2018. 4

    2018

  16. [24]

    Multipath affinage stacked—hourglass networks for human pose esti- mation

    Guoguang Hua, Lihong Li, and Shiguang Liu. Multipath affinage stacked—hourglass networks for human pose esti- mation. Frontiers of Computer Science, 14:1–12, 2020. 2

    2020

  17. [25]

    Ultralytics yolov5

    Glenn Jocher. Ultralytics yolov5. https://github.com/ ultra- lytics/yolov5/tree/v7.0, 2020. 2

    2020

  18. [26]

    Ultralytics yolov8

    Glenn Jocher. Ultralytics yolov8. https://github.com/ ultra- lytics/ultralytics/tree/v8.2.103, 2023. 1, 2

    2023

  19. [27]

    Ultralytics yolo11

    Glenn Jocher. Ultralytics yolo11. https://github.com/ ultra- lytics/ultralytics/tree/v8.3.27, 2024. 1, 2

    2024

  20. [28]

    Panoptic feature pyramid networks

    Alexander Kirillov, Ross Girshick, Kaiming He, and Piotr Doll´ar. Panoptic feature pyramid networks. In CVPR, pages 6399–6408, 2019. 2, 4, 1, 3

    2019

  21. [29]

    Pifpaf: Composite fields for human pose estimation

    Sven Kreiss, Lorenzo Bertoni, and Alexandre Alahi. Pifpaf: Composite fields for human pose estimation. InCVPR, pages 11977–11986, 2019. 1, 2

    2019

  22. [30]

    The hungarian method for the assignment problem

    Harold W Kuhn. The hungarian method for the assignment problem. Naval research logistics quarterly , 2(1-2):83–97,

  23. [31]

    Cornernet: Detecting objects as paired keypoints

    Hei Law and Jia Deng. Cornernet: Detecting objects as paired keypoints. In ECCV, pages 734–750, 2018. 1, 2, 4

    2018

  24. [32]

    Yolov6: A single-stage object detec- tion framework for industrial applications

    Chuyi Li, Lulu Li, Hongliang Jiang, Kaiheng Weng, Yifei Geng, Liang Li, Zaidan Ke, Qingyuan Li, Meng Cheng, Weiqiang Nie, et al. Yolov6: A single-stage object detec- tion framework for industrial applications. arXiv preprint arXiv:2209.02976, 2022. 2

    2022 arXiv

  25. [33]

    Generalized focal loss: Learning qualified and distributed bounding boxes for dense object detection

    Xiang Li, Wenhai Wang, Lijun Wu, Shuo Chen, Xiaolin Hu, Jun Li, Jinhui Tang, and Jian Yang. Generalized focal loss: Learning qualified and distributed bounding boxes for dense object detection. NeurIPS, 33:21002–21012, 2020. 4, 8

    2020

  26. [34]

    Simcc: A simple coordinate classification perspective for hu- man pose estimation

    Yanjie Li, Sen Yang, Peidong Liu, Shoukui Zhang, Yunx- iao Wang, Zhicheng Wang, Wankou Yang, and Shu-Tao Xia. Simcc: A simple coordinate classification perspective for hu- man pose estimation. In ECCV, pages 89–106. Springer,

  27. [35]

    Microsoft coco: Common objects in context

    Tsung-Yi Lin, Michael Maire, Serge Belongie, James Hays, Pietro Perona, Deva Ramanan, Piotr Doll´ar, and C Lawrence Zitnick. Microsoft coco: Common objects in context. In ECCV, pages 740–755. Springer, 2014. 2, 5

    2014

  28. [36]

    Feature pyramid networks for object detection

    Tsung-Yi Lin, Piotr Doll ´ar, Ross Girshick, Kaiming He, Bharath Hariharan, and Serge Belongie. Feature pyramid networks for object detection. In CVPR, pages 2117–2125,

  29. [37]

    Path aggregation network for instance segmentation

    Shu Liu, Lu Qi, Haifang Qin, Jianping Shi, and Jiaya Jia. Path aggregation network for instance segmentation. In CVPR, pages 8759–8768, 2018. 1, 3

    2018

  30. [38]

    Fully convolutional networks for semantic segmentation

    Jonathan Long, Evan Shelhamer, and Trevor Darrell. Fully convolutional networks for semantic segmentation. In CVPR, pages 3431–3440, 2015. 2, 3

    2015

  31. [39]

    Decoupled weight decay regularization

    I Loshchilov. Decoupled weight decay regularization. arXiv preprint arXiv:1711.05101, 2017. 6, 2

    2017 arXiv

  32. [40]

    Benchmarking ro- bustness in object detection: Autonomous driving when win- ter is coming

    Claudio Michaelis, Benjamin Mitzkus, Robert Geirhos, Evgenia Rusak, Oliver Bringmann, Alexander S Ecker, Matthias Bethge, and Wieland Brendel. Benchmarking ro- bustness in object detection: Autonomous driving when win- ter is coming. arXiv preprint arXiv:1907.07484, 2019. 1

    1907 arXiv

  33. [41]

    Associa- tive embedding: End-to-end learning for joint detection and grouping

    Alejandro Newell, Zhiao Huang, and Jia Deng. Associa- tive embedding: End-to-end learning for joint detection and grouping. NeurIPS, 30, 2017. 1, 2, 3

    2017

  34. [42]

    Person- lab: Person pose estimation and instance segmentation with a bottom-up, part-based, geometric embedding model

    George Papandreou, Tyler Zhu, Liang-Chieh Chen, Spyros Gidaris, Jonathan Tompson, and Kevin Murphy. Person- lab: Person pose estimation and instance segmentation with a bottom-up, part-based, geometric embedding model. In ECCV, pages 269–286, 2018. 3, 4

    2018

  35. [43]

    Mobilenetv4- universal models for the mobile ecosystem

    Danfeng Qin, Chas Leichner, Manolis Delakis, Marco Fornoni, Shixin Luo, Fan Yang, Weijun Wang, Colby Ban- bury, Chengxi Ye, Berkin Akin, et al. Mobilenetv4- universal models for the mobile ecosystem. arXiv preprint arXiv:2404.10518, 2024. 2, 4, 1

    2024 arXiv

  36. [44]

    Yolov3: An incremental improvement

    Joseph Redmon. Yolov3: An incremental improvement. arXiv preprint arXiv:1804.02767, 2018. 2

    2018 arXiv

  37. [45]

    Yolo9000: better, faster, stronger

    Joseph Redmon and Ali Farhadi. Yolo9000: better, faster, stronger. In CVPR, pages 7263–7271, 2017

    2017

  38. [46]

    You only look once: Unified, real-time object de- tection

    Joseph Redmon, Santosh Divvala, Ross Girshick, and Ali Farhadi. You only look once: Unified, real-time object de- tection. In CVPR, 2016. 2

    2016

  39. [47]

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

    Shaoqing Ren, Kaiming He, Ross Girshick, and Jian Sun. Faster r-cnn: Towards real-time object detection with region proposal networks. IEEE TPAMI, 39(6):1137–1149, 2016. 1

    2016

  40. [48]

    U- net: Convolutional networks for biomedical image segmen- tation

    Olaf Ronneberger, Philipp Fischer, and Thomas Brox. U- net: Convolutional networks for biomedical image segmen- tation. In Medical image computing and computer-assisted intervention–MICCAI 2015: 18th international conference, Munich, Germany, October 5-9, 2015, proceedings, pa...

    2015

  41. [49]

    Focal loss for dense object detection

    T-YLPG Ross and GKHP Doll ´ar. Focal loss for dense object detection. In CVPR, pages 2980–2988, 2017. 4, 8

    2017

  42. [50]

    Case study: Performance analysis and development of robotized screwing application with integrated vision sensing system for automotive indus- try

    Milan S ´aga, Vladim´ır Bulej, Nadeˇzda ˇCuboˇnova, Ivan Kuric, Ivan Virgala, and Manfred Eberth. Case study: Performance analysis and development of robotized screwing application with integrated vision sensing system for automotive indus- try. International Journal of Advanc...

    2020

  43. [51]

    Mobilenetv2: Inverted residuals and linear bottlenecks

    Mark Sandler, Andrew Howard, Menglong Zhu, Andrey Zh- moginov, and Liang-Chieh Chen. Mobilenetv2: Inverted residuals and linear bottlenecks. InCVPR, pages 4510–4520,

  44. [52]

    Objects365: A large-scale, high-quality dataset for object detection

    Shuai Shao, Zeming Li, Tianyuan Zhang, Chao Peng, Gang Yu, Xiangyu Zhang, Jing Li, and Jian Sun. Objects365: A large-scale, high-quality dataset for object detection. In ICCV, pages 8430–8439, 2019. 2

    2019

  45. [53]

    Intelligent video surveil- lance: a review through deep learning techniques for crowd analysis

    GSDMA Sreenu and Saleem Durai. Intelligent video surveil- lance: a review through deep learning techniques for crowd analysis. Journal of Big Data, 6(1):1–27, 2019. 1

    2019

  46. [54]

    Efficientnet: Rethinking model scaling for convolutional neural networks

    Mingxing Tan and Quoc Le. Efficientnet: Rethinking model scaling for convolutional neural networks. In International conference on machine learning (ICML), pages 6105–6114. PMLR, 2019. 3

    2019

  47. [55]

    Efficientnetv2: Smaller models and faster training

    Mingxing Tan and Quoc Le. Efficientnetv2: Smaller models and faster training. In International conference on machine learning (ICML), pages 10096–10106. PMLR, 2021. 2, 3

    2021

  48. [56]

    Fully convolutional one-stage 3d object detection on lidar range images

    Zhi Tian, Xiangxiang Chu, Xiaoming Wang, Xiaolin Wei, and Chunhua Shen. Fully convolutional one-stage 3d object detection on lidar range images. NeurIPS, 35:34899–34911,

  49. [57]

    Deeppose: Human pose estimation via deep neural networks

    Alexander Toshev and Christian Szegedy. Deeppose: Human pose estimation via deep neural networks. In CVPR, pages 1653–1660, 2014. 3

    2014

  50. [58]

    Mobileone: An improved one millisecond mobile backbone

    Pavan Kumar Anasosalu Vasu, James Gabriel, Jeff Zhu, On- cel Tuzel, and Anurag Ranjan. Mobileone: An improved one millisecond mobile backbone. In CVPR, pages 7907–7917,

  51. [59]

    Gomez, Łukasz Kaiser, and Illia Polosukhin

    Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszko- reit, Llion Jones, Aidan N. Gomez, Łukasz Kaiser, and Illia Polosukhin. Attention is all you need. NeurIPS, 2017. 4

    2017

  52. [60]

    Yolov10: Real-time end- to-end object detection

    Ao Wang, Hui Chen, Lihao Liu, Kai Chen, Zijia Lin, Jun- gong Han, and Guiguang Ding. Yolov10: Real-time end- to-end object detection. arXiv preprint arXiv:2405.14458 ,

    arXiv

  53. [61]

    Yolov7: Trainable bag-of-freebies sets new state-of-the-art for real-time object detectors

    Chien-Yao Wang, Alexey Bochkovskiy, and Hong- Yuan Mark Liao. Yolov7: Trainable bag-of-freebies sets new state-of-the-art for real-time object detectors. In CVPR, pages 7464–7475, 2023

    2023

  54. [62]

    Yolov9: Learning what you want to learn us- ing programmable gradient information

    Chien-Yao Wang, I-Hau Yeh, and Hong-Yuan Mark Liao. Yolov9: Learning what you want to learn us- ing programmable gradient information. arXiv preprint arXiv:2402.13616, 2024. 2

    2024 arXiv

  55. [63]

    Deep high-resolution represen- tation learning for visual recognition

    Jingdong Wang, Ke Sun, Tianheng Cheng, Borui Jiang, Chaorui Deng, Yang Zhao, Dong Liu, Yadong Mu, Mingkui Tan, Xinggang Wang, et al. Deep high-resolution represen- tation learning for visual recognition. IEEE TPAMI, 43(10): 3349–3364, 2020. 1, 2, 6, 3

    2020

  56. [64]

    Simple baselines for human pose estimation and tracking

    Bin Xiao, Haiping Wu, and Yichen Wei. Simple baselines for human pose estimation and tracking. In ECCV, pages 466–481, 2018. 1, 2, 3

    2018

  57. [65]

    Segformer: Simple and ef- ficient design for semantic segmentation with transformers

    Enze Xie, Wenhai Wang, Zhiding Yu, Anima Anandkumar, Jose M Alvarez, and Ping Luo. Segformer: Simple and ef- ficient design for semantic segmentation with transformers. NeurIPS, 34:12077–12090, 2021. 2, 3

    2021

  58. [66]

    Dart: An automated end-to-end object detection pipeline with data diversification, open-vocabulary bounding box annotation, pseudo-label review, and model training

    Chen Xin, Andreas Hartel, and Enkelejda Kasneci. Dart: An automated end-to-end object detection pipeline with data diversification, open-vocabulary bounding box annotation, pseudo-label review, and model training. Expert Systems with Applications, 258:125124, 2024. 2

    2024

  59. [67]

    Lite-hrnet: A lightweight high-resolution network

    Changqian Yu, Bin Xiao, Changxin Gao, Lu Yuan, Lei Zhang, Nong Sang, and Jingdong Wang. Lite-hrnet: A lightweight high-resolution network. In CVPR, pages 10440–10450, 2021. 2, 3

    2021

  60. [68]

    Pyramid scene parsing network

    Hengshuang Zhao, Jianping Shi, Xiaojuan Qi, Xiaogang Wang, and Jiaya Jia. Pyramid scene parsing network. In CVPR, pages 2881–2890, 2017. 2, 3

    2017

  61. [69]

    Detrs beat yolos on real-time object detection

    Yian Zhao, Wenyu Lv, Shangliang Xu, Jinman Wei, Guanzhong Wang, Qingqing Dang, Yi Liu, and Jie Chen. Detrs beat yolos on real-time object detection. In CVPR, pages 16965–16974, 2024. 1

    2024

  62. [70]

    Ob- jects as points

    Xingyi Zhou, Dequan Wang, and Philipp Kr ¨ahenb¨uhl. Ob- jects as points. arXiv preprint arXiv:1904.07850, 2019. 1, 2

    1904 arXiv

  63. [71]

    Learning-based object detection and localization for a mobile robot manipulator in sme production

    Zhengxue Zhou, Leihui Li, Alexander F ¨ursterling, Hjalte Joshua Durocher, Jesper Mouridsen, and Xup- ing Zhang. Learning-based object detection and localization for a mobile robot manipulator in sme production. Robotics and Computer-Integrated Manufacturing, 73:102229, 2022. 1

    2022

  64. [72]

    Detrs with collaborative hybrid assignments training

    Zhuofan Zong, Guanglu Song, and Yu Liu. Detrs with collaborative hybrid assignments training. In ICCV, pages 6748–6758, 2023. 1 OCDet: Object Center Detection via Bounding Box-Aware Heatmap Prediction on Edge Devices with NPUs Supplementary Material A. Figures Fig. 5 enlarges ...

    2023

  65. [73]

    This feature fusion utilizes standard Conv- BN-RelU and bilinear upsampling modules to construct a multi-scale feature pyramid (middle column)

    fuses features from the final stages of MobileNetV4 (left column). This feature fusion utilizes standard Conv- BN-RelU and bilinear upsampling modules to construct a multi-scale feature pyramid (middle column). OCDet mod- els leverage the last three or four levels of the pyram...

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.