Pith. sign in

REVIEW 5 major objections 6 minor 56 references

Dome-DETR: DETR with Density-Oriented Feature-Query Manipulation for Efficient Tiny Object Detection

T0 review · 5 major / 6 minor · reviewed 2026-08-15 · deepseek-v4-flash

Pith's one-line read Dome-DETR claims that a learned density map can simultaneously guide sparse attention and query allocation, pushing tiny-object AP to 34.6 on AI-TOD-V2 and 39.0 on VisDrone.

desk verdict Dome-DETR is a solid empirical step for tiny object detection built on D-FINE, with clean ablations, but its density-mask gating can silently drop isolated low-density objects and the SOTA framing outruns the evidence. read the letter →

arxiv 2505.05741 v2 pith:KTOYFQKA submitted 2025-05-09 cs.CV

classification cs.CV
keywords tinyobjectdetectiontransformerdensityestimationsparseattentionadaptivequeriesaerialimageryAI-TOD-V2VisDrone
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

Dome-DETR sets out to show that tiny-object detection in aerial imagery can be improved by letting a learned density map steer where the model spends computation and where it places its object queries. The paper claims this density-oriented manipulation pushes a D-FINE backbone to state-of-the-art results on two aerial benchmarks, with +3.3 AP on AI-TOD-V2 and +2.5 AP on VisDrone, while adding only modest parameters and FLOPs. A sympathetic reader should care because the approach directly attacks two known weaknesses of DETR-style detectors for small objects: shallow high-resolution features are expensive and diluted by background, and fixed query counts are wrong for scenes that range from one object to over a thousand. The paper argues that the same predicted density map can solve both problems at once.

What carries the argument

The load-bearing object is the predicted density map $D_{\text{pred}}$: DeFE turns the shallowest backbone feature map into a normalized heatmap using dilated depthwise convolutions and a density head, supervised by Density Recall Focal Loss. A thresholded binary mask $M_b$ derived from this map (Equations 5–7) chooses which windows MWAS keeps, and the same mask filters flexible queries in PAQI. MWAS then applies self-attention inside retained windows plus an axis-permuted second attention pass for cross-window communication; PAQI splits selected queries into a fixed core set and a density-filtered flexible set, then applies a density-dependent IoU threshold in NMS. The density map is the shared signal that ties feature sparsification to query allocation.

What would settle it

On AI-TOD-V2, compute the recall of the DeFE binary mask: count the fraction of ground-truth object centers that fall inside any retained window. If a substantial share of isolated small objects sits in pruned windows yet the reported AP still holds, the paper's explanation would be suspect; alternatively, ablating DeFE at inference by thresholding the density map at a fixed high value and observing a large AP drop would confirm that the mask, not the decoder, carries the gain.

Watch

Extended reading notes

Core claim

The central claim is that a single Density-Focal Extractor (DeFE) can produce a reliable pixel-level density heatmap of foreground objects from shallow backbone features, and that this heatmap can be reused to (i) prune background tokens from windowed self-attention via Masked Window Attention Sparsification (MWAS), and (ii) modulate the number, placement, and suppression threshold of decoder queries via Progressive Adaptive Query Initialization (PAQI). On the paper's own terms, the discovery is that these two uses are compatible and complementary: the density map focuses computation on informative windows while also giving dense scenes more queries and sparse scenes fewer, raising AP on AI-TOD-V2 from 30.1 to 33.3 for the small variant and reaching 34.6 AP for the large variant. The gains are largest for very tiny and tiny objects, which is the regime the components were designed for.

Load-bearing premise

The whole pipeline assumes the DeFE density prediction is reliable enough to gate both computation and queries: the binary mask built from it prunes any window with zero predicted density and discards candidate queries outside the mask, so objects the density estimator misses are structurally removed from the detection path.

Editorial extensions

If this is right

  • If the reported results hold, density-guided sparse attention can recover tiny-object accuracy lost by deep low-resolution features without paying global-attention cost on high-resolution maps.
  • Dynamic query allocation would remove the need to hand-tune query counts per dataset; the same model adapts from sparse scenes to scenes with over a thousand objects.
  • The computational cost of the detector scales with scene density rather than being fixed, so average GFLOPs understate cost in dense scenes and overstate it in sparse ones.
  • Because the components plug into a D-FINE-style encoder-decoder, the same density-guided manipulation could be transferred to other DETR variants with shallow features.

Reading between the lines

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

  • Editorial inference — a consequence the paper leaves implicit is that the three components are only as good as the density map: failures of DeFE would cascade into both MWAS and PAQI, so density-map recall should be reported alongside AP in follow-up work.
  • Editorial inference — the density heatmap could also serve other tasks such as object counting, saliency-guided crop selection, or adaptive video-frame sampling; the paper does not explore these uses.
  • Editorial inference — a testable extension is to train DeFE with DRFL but replace it at inference with ground-truth density maps; the performance gap would isolate how much of the gain comes from density-estimation accuracy versus the downstream gating mechanics.
  • Editorial inference — if DeFE is reliable, this line suggests that explicit density estimation may be a general cheap prior for DETR decoders in any image where foreground is sparse, not just aerial imagery.
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

5 major / 6 minor

Summary. Dome-DETR proposes a DETR-based tiny-object detector built on D-FINE, with three new components: a Density-Focal Extractor (DeFE) that predicts a density heatmap, Masked Window Attention Sparsification (MWAS) that prunes background windows using a thresholded density mask, and Progressive Adaptive Query Initialization (PAQI) that adaptively allocates object queries using the same density signal. The paper reports +3.3 AP on AI-TOD-V2 and +2.5 AP on VisDrone over D-FINE, claims state-of-the-art performance, and provides ablations and efficiency analyses. The code is publicly available.

Significance. If the reported gains hold, the paper is a useful empirical contribution to tiny-object detection: it shows that a lightweight density prior can jointly guide feature sparsification and query allocation, and it provides thorough experiments on two standard aerial benchmarks with clear component ablations. The authors also ship code and report a meaningful comparison against D-FINE, a strong recent baseline. However, the central efficiency-and-accuracy claim is currently undermined by an internal contradiction in the VisDrone state-of-the-art claim, an apparent inconsistency in the DeFE formulation, and the absence of object-level mask-recall analysis for the gating mechanism.

major comments (5)
  1. [Section 3.2, Eqs. (2)-(3)] As written, Eq. (2) defines F_G as the global average pooled feature over all spatial positions, and Eq. (3) computes D_pred by upsampling a function of F_G. Since F_G is spatially constant, Eq. (3) yields a constant density map, which contradicts the spatially varying density heatmaps shown in Figures 3 and 8 and the entire downstream gating mechanism. Please clarify the actual computation: if the density head operates on the spatially varying F'_S (perhaps with the GAP branch as an auxiliary channel-gate), rewrite Eqs. (2)-(3) accordingly; if the equations are correct, the model cannot produce the claimed masks.
  2. [Abstract, Section 1, Table 2, Appendix Table 8] The abstract and introduction claim state-of-the-art performance on VisDrone, but the paper's own Appendix Table 8 reports Focus&Detect at 42.0 AP and YOLC at 39.6 AP on the same validation split, both above Dome-DETR-L's 39.0 AP in Table 2. This is an internal contradiction in a load-bearing claim. Please either restrict the claim to DETR-based methods or to the latency-constrained setting, or provide a comparison that justifies the headline statement.
  3. [Section 3.3-3.4, Eqs. (5)-(8), Algorithm 1] The binary mask M_b is created by thresholding DeFE predictions, and both MWAS window pruning (Eq. 8) and PAQI flexible-query filtering (Algorithm 1, line 13) discard regions whose predicted density is at or below the threshold. The paper never measures the object-level recall of this mask, i.e., the fraction of ground-truth object centers covered by M_b or M_W, nor does it stratify failures by scene density. Because any object missed by DeFE is structurally removed from the feature-enhancement path and from the flexible query set, the reported AP gains cannot be causally attributed to density-oriented gating without this measurement. Please add mask-recall statistics and a failure analysis (e.g., recall at thresholds, recall by object size and by per-image object count).
  4. [Section 4.5, Table 1, Appendix Table 6] The GFLOPs comparison mixes static and data-dependent values: Dome-DETR and DQ-DETR are marked with a star for average GFLOPs, while D-FINE and other baselines report fixed single-image GFLOPs. Appendix Table 6 shows that in dense scenes Dome-DETR-L reaches 398.9 GFLOPs, which is 71.4 GFLOPs above D-FINE-L's 327.5, not 31.2 as suggested in Section 4.5. To support the 'low computational cost' claim, report worst-case or matched-protocol GFLOPs (or latency on identical hardware) for all compared methods, and state explicitly which numbers are averages and which are static.
  5. [Section 4.6, Table 3] The ablation study is reported from a single training run per configuration, and the per-component AP differences are 0.9-1.2 AP. On aerial benchmarks with single-seed evaluation, such differences can be within run-to-run variance. Please report the mean and standard deviation over at least three seeds, or provide statistical significance information, so that the claimed contributions of DeFE, MWAS, and PAQI are credible.
minor comments (6)
  1. [Section 4.4] The first sentence says 'Table 2 presents our main results on the AI-TOD-V2 test split,' but the AI-TOD-V2 results are in Table 1; the VisDrone sentence correctly refers to Table 2.
  2. [Section 4.6, Table 4] The hyperparameter paragraph says T_init=0.05 'achieves AP 32.6', but Table 4 reports 33.3 for that setting; the text should match the table.
  3. [Table 3] The header 'MS-WAS' should be 'MWAS' for consistency with the text and Figure 3.
  4. [Algorithm 1, Eq. (14)] The variable D in Eq. (14) is described as 'the predicted density of the final mask-filtered tokens,' but it is not clear whether T is per-query, per-window, or a single scalar per image; please define D and the scope of the IoU threshold computation.
  5. [Appendix E] The section titled 'Frequently Asked Questions During Review' is not a standard paper section; consider removing it or folding its content into the main text or a normal limitations/discussion section, especially the explicit references to reviewer suggestions.
  6. [Section 5] The conclusion repeats the phrase 'state-of-the-art results on the AI-TOD-V2 and VisDrone datasets'; this should be revised in light of the contradiction with Appendix Table 8.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: the method is supervised from ground-truth density, evaluated on external benchmarks, and the D-FINE self-citation is not load-bearing.

full rationale

Dome-DETR is an empirical detection framework rather than a derivation. DeFE is supervised by a ground-truth density map (Eq. 4), with D_gt built from Gaussian kernels centered on annotated objects, so the gating masks in Eqs. (5)-(8) and the query filtering in Algorithm 1 are not defined in terms of the final detection output. The central claims (+3.3 AP on AI-TOD-V2, +2.5 AP on VisDrone) are measured against external benchmark data and compared with published or re-implemented baselines; no equation reduces to its own input. The only authorship overlap is the D-FINE baseline [27], which is an independently published detector with public results; using it as the architectural base is a normal incremental-development choice, and the comparison is empirical rather than a citation carrying the argument. The QAR metric (Eq. 16) is self-defined but counts initialized queries versus ground-truth objects and is not used as evidence that the density-gating derivation is circular. Potential weaknesses, such as the absence of object-level mask recall for DeFE, are correctness or robustness concerns rather than circularity.

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

No new physical entities are hypothesized. DeFE, MWAS, and PAQI are software modules, not independent testable entities. All components are trained end-to-end on the benchmarks, so the only free parameters are architectural and threshold hyperparameters tuned on validation splits.

free parameters (4)
  • T_init (density threshold) = 0.05
    Initial threshold for the density mask in Eqs. (5) to (7); tuned in Table 4 and fixed in Table 9.
  • IoU_N / IoU_M (dynamic NMS endpoints) = 0.4 / 0.9
    Endpoints of the density-dependent NMS threshold in Eq. (14); ablated in Table 4.
  • MWAS window size (H/W) = 10
    Window size for attention sparsification; ablated in Table 4.
  • Max/Min query counts (K_M, K_N) = 1500 / 300
    Query budget for PAQI in Algorithm 1 and Appendix B; dense-scene maximum and sparse-scene minimum.
assumptions (4)
  • domain assumption Pretrained ImageNet backbone features transfer to aerial tiny objects.
    All variants use HGNetV2 pretrained on ImageNet, per Table 9.
  • domain assumption Gaussian-blurred ground truth is a valid density supervision signal.
    The density target is generated by convolving Gaussian kernels over object centers, Section 3.2.
  • domain assumption COCO-style AP with a 1500-detection limit is the accepted evaluation for these benchmarks.
    Used for all reported scores; only the appendix reports the auxiliary LRP metric.
  • domain assumption The D-FINE baseline and its re-implementations are faithful.
    Dome-DETR is built on D-FINE and compares against re-implemented D-FINE results; authors overlap with D-FINE's authors.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Dome-DETR: DETR with Density-Oriented Feature-Query Manipulation for Efficient Tiny Object Detection." pith.science (2026). https://pith.science/paper/KTOYFQKA

@misc{pith2026250505741,
  author       = {Pith},
  title        = {Pith review of: Dome-DETR: DETR with Density-Oriented Feature-Query Manipulation for Efficient Tiny Object Detection},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/KTOYFQKA}},
  note         = {Machine review of arXiv:2505.05741}
}
read the original abstract

Tiny object detection plays a vital role in drone surveillance, remote sensing, and autonomous systems, enabling the identification of small targets across vast landscapes. However, existing methods suffer from inefficient feature leverage and high computational costs due to redundant feature processing and rigid query allocation. To address these challenges, we propose Dome-DETR, a novel framework with Density-Oriented Feature-Query Manipulation for Efficient Tiny Object Detection. To reduce feature redundancies, we introduce a lightweight Density-Focal Extractor (DeFE) to produce clustered compact foreground masks. Leveraging these masks, we incorporate Masked Window Attention Sparsification (MWAS) to focus computational resources on the most informative regions via sparse attention. Besides, we propose Progressive Adaptive Query Initialization (PAQI), which adaptively modulates query density across spatial areas for better query allocation. Extensive experiments demonstrate that Dome-DETR achieves state-of-the-art performance (+3.3 AP on AI-TOD-V2 and +2.5 AP on VisDrone) while maintaining low computational complexity and a compact model size. Code is available at https://github.com/RicePasteM/Dome-DETR.

Figures

Figures reproduced from arXiv: 2505.05741 by the authors.

Figure 1
Figure 1. Comparisons with other detectors on the AI-TOD [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 3
Figure 3. The overall pipeline of our proposed model. The process begins with the backbone network extracting multi-scale [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figure 4
Figure 4. An illustration of the (A) Axis Permuted Encoder [PITH_FULL_IMAGE:figures/full_fig_p005_4.png] view at source ↗
Figures from the paper (4 more)
Figure 5
Figure 5. Figure 5: Visualization of comparison between our proposed method and other methods under different scenes on AI-TOD-V2 [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]
Figure 6
Figure 6. Figure 6: Ground truth counts and corresponding enhanced [PITH_FULL_IMAGE:figures/full_fig_p012_6.png]
Figure 7
Figure 7. Figure 7: Ground truth counts and corresponding query [PITH_FULL_IMAGE:figures/full_fig_p012_7.png]
Figure 8
Figure 8. Figure 8: Visualization of our proposed method under different scenes across two datasets. [PITH_FULL_IMAGE:figures/full_fig_p014_8.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

56 extracted references · 26 canonical work pages

  1. [1]

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

  2. [2]

    Zhaowei Cai and Nuno Vasconcelos. 2018. Cascade r-cnn: Delving into high quality object detection. In Proceedings of the IEEE conference on computer vision and pattern recognition. 6154–6162

  3. [3]

    Nicolas Carion, Francisco Massa, Gabriel Synnaeve, Nicolas Usunier, Alexan- der Kirillov, and Sergey Zagoruyko. 2020. End-to-end object detection with transformers. In European conference on computer vision . Springer, 213–229

  4. [4]

    Qiang Chen, Xiangbo Su, Xinyu Zhang, Jian Wang, Jiahui Chen, Yunpeng Shen, Chuchu Han, Ziliang Chen, Weixiang Xu, Fanrong Li, et al . 2024. LW- DETR: a transformer replacement to yolo for real-time detection. arXiv preprint arXiv:2406.03459 (2024)

  5. [5]

    Jifeng Dai, Yi Li, Kaiming He, and Jian Sun. 2016. R-fcn: Object detection via region-based fully convolutional networks. Advances in neural information processing systems 29 (2016)

  6. [6]

    Bowei Du, Yecheng Huang, Jiaxin Chen, and Di Huang. 2023. Adaptive sparse convolutional networks with global context enhancement for faster object de- tection on drone images. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition . 13435–13444

  7. [7]

    Ross Girshick. 2015. Fast r-cnn. In Proceedings of the IEEE international conference on computer vision. 1440–1448

  8. [8]

    Kaiming He, Georgia Gkioxari, Piotr Dollár, and Ross Girshick. 2017. Mask r-cnn. In Proceedings of the IEEE international conference on computer vision . 2961–2969

Show all 56 references
  1. [9]

    Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. 2016. Deep residual learning for image recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition . 770–778

  2. [10]

    Shihua Huang, Zhichao Lu, Xiaodong Cun, Yongjun Yu, Xiao Zhou, and Xi Shen

  3. [11]

    Yecheng Huang, Jiaxin Chen, and Di Huang. 2022. UFPMP-Det: Toward accurate and efficient object detection on drone imagery. In Proceedings of the AAAI conference on artificial intelligence , Vol. 36. 1026–1033

  4. [12]

    Yi-Xin Huang, Hou-I Liu, Hong-Han Shuai, and Wen-Huang Cheng. 2024. Dq- detr: Detr with dynamic query for tiny object detection. In European Conference on Computer Vision. Springer, 290–305

  5. [13]

    Mate Kisantal, Zbigniew Wojna, Jakub Murawski, Jacek Naruniec, and Kyunghyun Cho. 2019. Augmentation for small object detection. arXiv preprint arXiv:1902.07296 (2019)

  6. [14]

    Onur Can Koyun, Reyhan Kevser Keser, Ibrahim Batuhan Akkaya, and Be- hçet Uğur Töreyin. 2022. Focus-and-Detect: A small object detection framework for aerial images. Signal Processing: Image Communication 104 (2022), 116675

  7. [15]

    Feng Li, Hao Zhang, Shilong Liu, Jian Guo, Lionel M Ni, and Lei Zhang. 2022. Dn- detr: Accelerate detr training by introducing query denoising. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition . 13619–13627

  8. [16]

    Tsung-Yi Lin, Piotr Dollár, Ross Girshick, Kaiming He, Bharath Hariharan, and Serge Belongie. 2017. Feature pyramid networks for object detection. In Proceed- ings of the IEEE conference on computer vision and pattern recognition . 2117–2125

  9. [17]

    Chenguang Liu, Guangshuai Gao, Ziyue Huang, Zhenghui Hu, Qingjie Liu, and Yunhong Wang. 2024. Yolc: You only look clusters for tiny object detection in aerial images. IEEE transactions on intelligent transportation systems 25, 10 (2024), 13863–13875

  10. [18]

    Hou-I Liu, Yu-Wen Tseng, Kai-Cheng Chang, Pin-Jyun Wang, Hong-Han Shuai, and Wen-Huang Cheng. 2024. A denoising fpn with transformer r-cnn for tiny object detection. IEEE Transactions on Geoscience and Remote Sensing (2024)

  11. [19]

    Shilong Liu, Feng Li, Hao Zhang, Xiao Yang, Xianbiao Qi, Hang Su, Jun Zhu, and Lei Zhang. 2022. Dab-detr: Dynamic anchor boxes are better queries for detr. arXiv preprint arXiv:2201.12329 (2022)

  12. [20]

    Wei Liu, Dragomir Anguelov, Dumitru Erhan, Christian Szegedy, Scott Reed, Cheng-Yang Fu, and Alexander C Berg. 2016. Ssd: Single shot multibox detec- tor. In Computer Vision–ECCV 2016: 14th European Conference, Amsterdam, The Netherlands, October 11–14, 2016, Proceedings, Part...

  13. [21]

    Yang Liu, Peng Sun, Nickolas Wergeles, and Yi Shang. 2021. A survey and performance evaluation of deep learning methods for small object detection. Expert Systems with Applications 172 (2021), 114602

  14. [22]

    Depu Meng, Xiaokang Chen, Zejia Fan, Gang Zeng, Houqiang Li, Yuhui Yuan, Lei Sun, and Jingdong Wang. 2021. Conditional detr for fast training convergence. In Proceedings of the IEEE/CVF international conference on computer vision . 3651– 3660

  15. [23]

    Payal Mittal, Raman Singh, and Akashdeep Sharma. 2020. Deep learning-based object detection in low-altitude UAV datasets: A survey. Image and Vision com- puting 104 (2020), 104046

  16. [24]

    MiXaiLL76. 2024. Faster-COCO-Eval: Faster interpretation of the original CO- COEval. (2024)

  17. [25]

    Jinlai Ning, Haoyan Guan, and Michael Spratling. 2023. Rethinking the backbone architecture for tiny object detection. arXiv preprint arXiv:2303.11267 (2023)

  18. [26]

    Kemal Oksuz, Baris Can Cam, Emre Akbas, and Sinan Kalkan. 2020. A Ranking- based, Balanced Loss Function Unifying Classification and Localisation in Object Detection. arXiv:arXiv:2009.13592

  19. [27]

    Yansong Peng, Hebei Li, Peixi Wu, Yueyi Zhang, Xiaoyan Sun, and Feng Wu

  20. [28]

    Yansong Peng, Hebei Li, Yueyi Zhang, Xiaoyan Sun, and Feng Wu. 2024. Scene Adaptive Sparse Transformer for Event-based Object Detection. In2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) . 16794–16804. https://doi.org/10.1109/CVPR52733.2024.01589

  21. [29]

    arXiv preprint arXiv:2410.13842 (2024)

    D-FINE: redefine regression Task in DETRs as Fine-grained distribution refinement. arXiv preprint arXiv:2410.13842 (2024)

  22. [30]

    Joseph Redmon, Santosh Divvala, Ross Girshick, and Ali Farhadi. 2016. You only look once: Unified, real-time object detection. In Proceedings of the IEEE conference on computer vision and pattern recognition . 779–788

  23. [31]

    Siyuan Qiao, Liang-Chieh Chen, and Alan Yuille. 2021. Detectors: Detecting objects with recursive feature pyramid and switchable atrous convolution. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition . 10213–10224

  24. [32]

    Chao Shen, Caiwen Ma, and Wei Gao. 2023. Multiple attention mechanism enhanced YOLOX for remote sensing object detection. Sensors 23, 3 (2023), 1261

  25. [33]

    Shaoqing Ren, Kaiming He, Ross Girshick, and Jian Sun. 2016. Faster R-CNN: To- wards real-time object detection with region proposal networks.IEEE transactions on pattern analysis and machine intelligence 39, 6 (2016), 1137–1149

  26. [34]

    Yunjie Tian, Qixiang Ye, and David Doermann. 2025. Yolov12: Attention-centric real-time object detectors. arXiv preprint arXiv:2502.12524 (2025)

  27. [35]

    Shiyi Tang, Shu Zhang, and Yini Fang. 2024. HIC-YOLOv5: Improved YOLOv5 for small object detection. In 2024 IEEE International Conference on Robotics and Automation (ICRA). IEEE, 6614–6619

  28. [36]

    Zhengzhong Tu, Hossein Talebi, Han Zhang, Feng Yang, Peyman Milanfar, Alan Bovik, and Yinxiao Li. 2022. Maxvit: Multi-axis vision transformer. In European conference on computer vision . Springer, 459–479

  29. [37]

    Zhi Tian, Chunhua Shen, Hao Chen, and Tong He. 2019. Fcos: Fully convolutional one-stage object detection. In Proceedings of the IEEE/CVF international conference on computer vision. 9627–9636

  30. [38]

    Chengcheng Wang, Wei He, Ying Nie, Jianyuan Guo, Chuanjian Liu, Yunhe Wang, and Kai Han. 2023. Gold-YOLO: Efficient object detector via gather-and- distribute mechanism. Advances in Neural Information Processing Systems 36 (2023), 51094–51112

  31. [39]

    Ao Wang, Hui Chen, Lihao Liu, Kai Chen, Zijia Lin, Jungong Han, et al . 2024. Yolov10: Real-time end-to-end object detection. Advances in Neural Information Processing Systems 37 (2024), 107984–108011

  32. [40]

    Jinwang Wang, Wen Yang, Haowen Guo, Ruixiang Zhang, and Gui-Song Xia

  33. [41]

    Jinwang Wang, Chang Xu, Wen Yang, and Lei Yu. 2021. A normalized Gaussian Wasserstein distance for tiny object detection. arXiv preprint arXiv:2110.13389 (2021)

  34. [42]

    Chang Xu, Jinwang Wang, Wen Yang, Huai Yu, Lei Yu, and Gui-Song Xia. 2022. RFLA: Gaussian receptive field based label assignment for tiny object detection. In European conference on computer vision . Springer, 526–543

  35. [43]

    Chang Xu, Jinwang Wang, Wen Yang, and Lei Yu. 2021. Dot distance for tiny object detection in aerial images. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition . 1192–1201

  36. [44]

    Chang Xu, Jinwang Wang, Wen Yang, Huai Yu, Lei Yu, and Gui-Song Xia. 2022. Detecting tiny objects in aerial images: A normalized Wasserstein distance and a new benchmark. ISPRS Journal of Photogrammetry and Remote Sensing 190 (2022), 79–93

  37. [45]

    Fan Yang, Heng Fan, Peng Chu, Erik Blasch, and Haibin Ling. 2019. Clustered object detection in aerial images. In Proceedings of the IEEE/CVF international conference on computer vision . 8311–8320

  38. [46]

    Hongyi Zhang, Moustapha Cisse, Yann N Dauphin, and David Lopez-Paz. 2017. mixup: Beyond empirical risk minimization. arXiv preprint arXiv:1710.09412 (2017)

  39. [47]

    Chenhongyi Yang, Zehao Huang, and Naiyan Wang. 2022. QueryDet: Cascaded sparse query for accelerating high-resolution small object detection. In Pro- ceedings of the IEEE/CVF Conference on computer vision and pattern recognition . 13668–13677

  40. [48]

    Huaxiang Zhang, Kai Liu, Zhongxue Gan, and Guo-Niu Zhu. 2025. UAV-DETR: Efficient End-to-End Object Detection for Unmanned Aerial Vehicle Imagery. MM ’25, October 27–31, 2025, Dublin, Ireland. Zhangchi Hu et al. arXiv preprint arXiv:2501.01855 (2025)

  41. [49]

    Shilong Zhang, Xinjiang Wang, Jiaqi Wang, Jiangmiao Pang, Chengqi Lyu, Wen- wei Zhang, Ping Luo, and Kai Chen. 2023. Dense distinct query for end-to-end object detection. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. 7329–7338

  42. [50]

    Hao Zhang, Feng Li, Shilong Liu, Lei Zhang, Hang Su, Jun Zhu, Lionel M Ni, and Heung-Yeung Shum. 2022. Dino: Detr with improved denoising anchor boxes for end-to-end object detection. arXiv preprint arXiv:2203.03605 (2022)

  43. [51]

    Yian Zhao, Wenyu Lv, Shangliang Xu, Jinman Wei, Guanzhong Wang, Qingqing Dang, Yi Liu, and Jie Chen. 2024. Detrs beat yolos on real-time object detection. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition . 16965–16974

  44. [52]

    Pengfei Zhu, Longyin Wen, Dawei Du, Xiao Bian, Heng Fan, Qinghua Hu, and Haibin Ling. 2021. Detection and tracking meet drones challenge. IEEE Transac- tions on Pattern Analysis and Machine Intelligence 44, 11 (2021), 7380–7399

  45. [53]

    ZhangchiHu. 2025. Faster-COCO-Eval-AITOD: Faster interpretation of the origi- nal aitodpycocotools. (2025)

  46. [56]

    spatial atten- tion guide

    Xizhou Zhu, Weijie Su, Lewei Lu, Bin Li, Xiaogang Wang, and Jifeng Dai. 2020. Deformable detr: Deformable transformers for end-to-end object detection.arXiv preprint arXiv:2010.04159 (2020). Dome-DETR: DETR with Density-Oriented Feature-Query Manipulation for Efficient Tiny Ob...

  47. [2021]

    In 2020 25th international conference on pattern recognition (ICPR)

    Tiny object detection in aerial images. In 2020 25th international conference on pattern recognition (ICPR) . IEEE, 3791–3798

  48. [2024]

    DEIM: DETR with Improved Matching for Fast Convergence.arXiv preprint arXiv:2412.04234 (2024)

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.