Pith. sign in

REVIEW 4 major objections 5 minor 55 references

Delving into Robust Object Detection from Unmanned Aerial Vehicles: A Deep Nuisance Disentanglement Approach

T0 review · 4 major / 5 minor · reviewed 2026-08-14 · deepseek-v4-flash

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

Pith's one-line read This paper claims that an adversarial feature transform suppressing altitude, view-angle, and weather information in UAV images produces substantially more robust object detectors.

desk verdict A practical UAV robustness method with consistent gains, but the 'disentanglement' story is unverified—worth reviewing on the strength of the applied result. read the letter →

arxiv 1908.03856 v2 pith:MPLC2WO6 submitted 2019-08-11 cs.CV

Zhenyu Wu , Karthik Suresh , Priya Narayanan , Hongyu Xu , Heesung Kwon , Zhangyang Wang This is my paper · ORCID
classification cs.CV
keywords objectdetectionunmannedaerialvehiclesadversarialtrainingfeaturedisentanglementnuisancemetadatanegativeentropylossdomainrobustnesscross-domaintransfer
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

The paper sets out to show that UAV-specific visual nuisances—altitude, camera view angle, and weather—can be treated as adversarial signals rather than static domain shifts, and that removing them from the learned features improves object detection substantially. It proposes Nuisance Disentangled Feature Transform (NDFT), a feature-extraction module trained in a three-way game: a detector must still recognize vehicles from the transformed features, while separate classifiers trained on UAV metadata must be unable to read those nuisance attributes from the same features. On the UAVDT benchmark, NDFT raises average precision from 45.64 to 47.91; on VisDrone2018 it raises mAP from 48.41 to 52.77, and the learned features transfer to a new dataset better than features trained without disentanglement. The practical payoff is that the nuisance labels are essentially free, since altitude comes from onboard sensors and weather from flight time and location, so the robustness gain does not require extra annotation effort.

What carries the argument

The load-bearing object is the Nuisance Disentangled Feature Transform $f_T$, implemented as the early convolutional layers of the detector (conv1–conv4 in the Faster-RCNN ResNet-101 example) and shared by both downstream branches. The mechanism is the three-party game: the detector $f_O$ and transform $f_T$ jointly minimize the detection loss plus weighted negative-entropy terms $\mathcal{L}_{ne}(f^i_N(f_T(X)))$ for each nuisance, while each nuisance classifier $f^i_N$ is alternately trained on labeled metadata to keep predicting altitude, view, and weather accurately. Minimizing negative entropy pushes the nuisance classifiers toward uniform, uncertain predictions, which is the operational definition of “nuisance-free” features used in the paper. Algorithm 1 stabilizes the game by refreshing the nuisance classifiers when their training accuracy falls below 0.9 and re-initializing them every 1000 iterations, so that $f_T$ cannot win by merely fooling a weak adversary.

What would settle it

Train NDFT with the true nuisance labels replaced by randomly shuffled labels (same class distribution, no correspondence to images) and compare detection AP. If the AP gain over the baseline largely persists, the improvement is a regularization effect and the disentanglement story is false; alternatively, train the three nuisance classifiers on the final frozen $f_T$ features and check whether their accuracy and entropy are actually worse than on the baseline features.

Watch

Extended reading notes

Core claim

The central claim is that one can learn a feature transform $f_T$ that keeps the information needed for object detection while discarding the information behind UAV-specific nuisances, and that doing so yields detectors that generalize across the many fine-grained domains created by altitude, view, and weather combinations. The authors formulate this as an alternating adversarial objective: $\min_{f_O,f_T} \mathcal{L}_O(f_O(f_T(X)),Y_O) + \sum_i \gamma_i \mathcal{L}_{ne}(f^i_N(f_T(X)))$, while the nuisance predictors $f^i_N$ are trained with their classification losses to remain strong adversaries. Here $\mathcal{L}_{ne}$ is the negative entropy of each nuisance predictor's output distribution, so the transform is rewarded for making altitude, view, and weather predictions uniformly uncertain. The paper reports that on UAVDT the full A+V+W model reaches 47.91 AP versus a 45.64 baseline; on VisDrone2018 NDFT-DE-FPN reaches 52.77 mAP versus 48.41 for DE-FPN; and a transferred NDFT feature extractor reaches 79.50 AP on vehicles versus 75.27 for the non-disentangled transfer baseline. The authors read these gains as evidence that explicit nuisance disentanglement, rather than auxiliary multi-task learning, is what drives robustness.

Load-bearing premise

The paper assumes that the adversarial loss that pushes nuisance predictions toward uncertainty genuinely removes altitude, view-angle, and weather information from the learned features rather than merely acting as a regularizer, but it never directly measures whether those predictions actually become less accurate on the final features.

Editorial extensions

If this is right

  • Fusing NDFT into Faster-RCNN raises UAVDT overall AP from 45.64 to 47.91, with the best result when all three nuisances are disentangled together.
  • NDFT is backbone-agnostic: with an FPN backbone on UAVDT it lifts AP from 49.05 to 52.03, and on VisDrone2018 it lifts mAP from 48.41 to 52.77 over the strong DE-FPN baseline.
  • NDFT features transfer across datasets: a DE-FPN head trained on UAVDT with NDFT features reaches 79.50 AP for vehicles on VisDrone2018, compared to 75.27 for the non-disentangled transfer baseline.
  • Using the nuisance labels adversarially outperforms using the same labels as auxiliary multi-task losses, which gives only a marginal gain over the baseline.
  • The improved detections also propagate to multi-object tracking: NDFT detections improve SORT tracker performance on 10 of 11 UAVDT MOT metrics.

Reading between the lines

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

  • A natural extension is to apply NDFT to automatically discovered nuisance groupings instead of the three hand-specified labels, which would test whether the mechanism generalizes to nuisance types that are not annotated.
  • Because the gains concentrate in the hardest nuisance classes (night, bird-view, high altitude), deploying NDFT could be most valuable in adverse operating conditions, and future work could quantify robustness on unseen nuisance levels rather than only held-out classes.
  • The same three-party game formulation used for privacy-preserving recognition suggests the transform could be tuned to suppress private attributes while preserving detection utility; the paper notes the connection but leaves the cross-task transfer untested.
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

4 major / 5 minor

Summary. The paper proposes Nuisance Disentangled Feature Transform (NDFT), an adversarial training framework for object detection in UAV imagery. The central idea is to learn a feature transform f_T that preserves object-detection information while suppressing UAV-specific nuisance information (altitude, view angle, weather) by training nuisance classifiers f_N adversarially. The authors implement NDFT on Faster-RCNN and DE-FPN backbones and report consistent AP/mAP gains on the UAVDT and VisDrone2018 datasets, including a transfer experiment from UAVDT to VisDrone2018. The paper also reports an ablation study, a multi-task learning comparison, and a tracking proof-of-concept.

Significance. If the empirical gains are genuine and the mechanism is as claimed, NDFT would be a practically valuable method for UAV object detection, leveraging freely available metadata. The consistent gains across two datasets, two backbones, and a held-out transfer setting are encouraging, and the released code supports reproducibility. However, the paper's central explanatory claim—that the method achieves 'nuisance disentanglement'—is never directly verified, and the reported gains are selected from evaluation-set grid searches without error bars. The transfer experiment provides the most credible evidence of generalization, but the overall significance is tempered by these methodological gaps.

major comments (4)
  1. [Section 3.2, Eq. (3), Algorithm 1] The implemented adversarial loss is the negative entropy L_ne of the current nuisance classifier's softmax outputs, not the minimax objective in Eq. (1)/(2). The paper never directly measures whether f_T actually suppresses nuisance information: no nuisance-prediction accuracy, prediction entropy, or information-theoretic estimate on the final f_T features is reported. The only evidence is downstream detection AP. This is load-bearing because minimizing L_ne can be achieved by making the particular f_N uncertain (e.g., shrinking logits) without making features statistically independent of the nuisance labels. Please add direct measurements of nuisance information in f_T (e.g., held-out nuisance classification accuracy or entropy) and relate them to the detection gains; otherwise the 'disentanglement' explanation is unsupported and the gains could be due to regularization.
  2. [Tables 1-3 and Table 7] The gamma coefficients are selected by grid search directly on the evaluation sets (UAVDT test set and VisDrone2018 validation set), and no repeated runs or error bars are reported. This makes it impossible to judge whether the reported gains are statistically meaningful or artifacts of tuning. For example, the altitude-only gain in Table 1 is 0.28 AP, smaller than likely run-to-run variance in deep detection training. Please provide error bars over multiple runs, select hyperparameters on a validation split that is disjoint from the reported test set, or justify why the selected gamma values are not overfit to the evaluation sets.
  3. [Tables 1, 3, and 4] The class-wise results show trade-offs that are in tension with the claim of clean nuisance disentanglement. In Table 1, the altitude-only model (gamma1=0.03) reduces high-altitude AP from 18.70 to 15.69 while improving medium-altitude AP; in Table 3, the weather-only model (gamma3=0.01) reduces day AP from 45.63 to 45.18 while improving night AP by 7.52; and in Table 4, several nuisance class cells (e.g., low altitude for A+W, front view for A+W) decrease relative to baseline. This pattern suggests that NDFT is not simply removing nuisance information without affecting task-relevant features, but rather re-balancing performance across nuisance conditions. Please discuss these trade-offs explicitly and provide a measure of whether the entropy loss actually reduces nuisance predictability for the classes that show degradation.
  4. [Algorithm 1] The training algorithm contains two ad-hoc components that are not ablated: the 'strength' criterion based on per-mini-batch training accuracy (threshold 0.9) and the periodic random re-initialization of all f_N branches every 1000 iterations. The non-stationarity introduced by the restarts means the adversary is not a consistent signal, and the threshold criterion does not measure generalization of the adversary. The paper claims these tricks help escape bad local minima, but provides no ablation or analysis. Please add an ablation study for these two components, or at least a sensitivity analysis, to show that the reported results do not depend critically on these unstated choices.
minor comments (5)
  1. [Section 4.3] The text reads 'receptively' and should be 'respectively'. Please also consider clarifying that the transfer experiment fine-tunes only the classification/regression head while freezing f_T, since this is a strong condition that may affect the interpretation.
  2. [Section 3.2] The phrase 'we will keep monitering f_N branches' contains a typo ('monitering' should be 'monitoring'). More substantively, the description of when nuisance predictor updates are triggered ('if at least one f_N becomes too weak') is ambiguous: the pseudocode uses a while-loop condition on training accuracy, but the prose suggests a per-mini-batch check. Please align these descriptions.
  3. [Tables 1-4] The tables mix rows for nuisance classes and overall AP; consider adding a separator or a note to distinguish the 'Overall' row from class-wise rows. In Table 5, the metric names are run together with values; please present this as a table with clear columns.
  4. [References] Some references are incomplete or have inconsistent formatting (e.g., [4] is a URL without a year; [28] lacks a venue; [37] has no full citation). Please ensure all references are complete and consistently formatted.
  5. [Section 4.2] The paper states that the nuisance annotations for VisDrone2018 are manually annotated and will be released. Please clarify the annotation protocol (e.g., how weather and altitude are determined from static images) and whether the annotations are per-image or per-sequence, as this affects reproducibility.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity; the central claims are empirical and externally benchmarked, with self-citations only in supporting roles.

full rationale

NDFT is an empirical method, not a derivation that could collapse into its own inputs. Equation (3) defines the training objective (detection loss L_O plus negative-entropy nuisance losses L_ne) and Algorithm 1 specifies the alternating adversarial updates. The paper's central evidence is external: UAVDT AP (Table 4), VisDrone mAP (Table 7), and the UAVDT-to-VisDrone transfer vehicle AP (Section 4.3). These are held-out benchmark numbers, not identities implied by the objective. The gamma coefficients are tuned hyperparameters selected on the evaluation sets; this is a fitting/selection issue, not a fitted input renamed as a prediction. The self-citations [45] and [48] support related work and a training restart heuristic, and are not used as an external uniqueness theorem or as the sole justification for the central result. The notable gap is that nuisance suppression in f_T is never directly measured (only downstream AP), so the 'disentanglement' mechanism is not verified; however, this is an empirical-validation weakness, not circular reasoning. No equation in the paper reduces by construction to its own input.

Assumptions & free parameters 5 free parameters · 4 assumptions · 0 invented entities

The central claim rests on the correctness of the adversarial disentanglement objective, the reliability of the nuisance labels, and several training heuristics. The gamma coefficients are the main fitted parameters; the monitoring threshold and restart period are hand-chosen training tricks. No new physical or conceptual entities are introduced.

free parameters (5)
  • gamma_1 (altitude adversarial weight) = 0.01 for UAVDT Faster-RCNN; 0.005 for UAVDT FPN and VisDrone DE-FPN
    Chosen by grid search over values such as 0, 0.01, 0.02, 0.03, 0.05 to maximize validation AP, then the best value is reported.
  • gamma_2 (view angle adversarial weight) = 0.01 for UAVDT Faster-RCNN; 0.005 for UAVDT FPN and VisDrone DE-FPN
    Selected by grid search and reported at the best-performing value on the evaluation set.
  • gamma_3 (weather adversarial weight) = 0.01 for UAVDT Faster-RCNN; 0.005 for UAVDT FPN and VisDrone DE-FPN
    Selected by grid search and reported at the best-performing value on the evaluation set.
  • nuisance classifier accuracy monitoring threshold = 0.9
    Hand-chosen threshold in Algorithm 1 to decide when to re-train the nuisance classifiers.
  • nuisance classifier restart period = 1000 iterations
    Hand-chosen period for re-initializing nuisance classifiers, described as an empirical trick to avoid bad local minima.
assumptions (4)
  • domain assumption Stochastic gradient descent on the adversarial objective converges to a feature transform that preserves detection information while suppressing nuisance information.
    Algorithm 1 and Section 3.2 rely on this convergence without proof or verification of the final nuisance prediction accuracy.
  • domain assumption The metadata labels (altitude, view, weather) are accurate and capture the dominant nuisance factors in UAV imagery.
    Section 4.1 and 4.2 use these labels as ground truth; on VisDrone they are manually annotated with no reported inter-annotator agreement.
  • domain assumption Minimizing the negative entropy of nuisance predictions removes nuisance-related information from f_T(X).
    Section 3.1, Eq. (3), states that L_ne encourages uncertain predictions, but the paper does not measure entropy reduction or nuisance information removal directly.
  • ad hoc to paper Periodically re-initializing the nuisance classifiers and re-training them helps the adversarial training escape bad local minima.
    Section 3.2 describes this restart trick as empirically beneficial, with no theoretical justification.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Delving into Robust Object Detection from Unmanned Aerial Vehicles: A Deep Nuisance Disentanglement Approach." pith.science (2026). https://pith.science/paper/MPLC2WO6

@misc{pith2026190803856,
  author       = {Pith},
  title        = {Pith review of: Delving into Robust Object Detection from Unmanned Aerial Vehicles: A Deep Nuisance Disentanglement Approach},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/MPLC2WO6}},
  note         = {Machine review of arXiv:1908.03856}
}
read the original abstract

Object detection from images captured by Unmanned Aerial Vehicles (UAVs) is becoming increasingly useful. Despite the great success of the generic object detection methods trained on ground-to-ground images, a huge performance drop is observed when they are directly applied to images captured by UAVs. The unsatisfactory performance is owing to many UAV-specific nuisances, such as varying flying altitudes, adverse weather conditions, dynamically changing viewing angles, etc. Those nuisances constitute a large number of fine-grained domains, across which the detection model has to stay robust. Fortunately, UAVs will record meta-data that depict those varying attributes, which are either freely available along with the UAV images, or can be easily obtained. We propose to utilize those free meta-data in conjunction with associated UAV images to learn domain-robust features via an adversarial training framework dubbed Nuisance Disentangled Feature Transform (NDFT), for the specific challenging problem of object detection in UAV images, achieving a substantial gain in robustness to those nuisances. We demonstrate the effectiveness of our proposed algorithm, by showing state-of-the-art performance (single model) on two existing UAV-based object detection benchmarks. The code is available at https://github.com/TAMU-VITA/UAV-NDFT.

Figures

Figures reproduced from arXiv: 1908.03856 by the authors.

Figure 1
Figure 1. Examples showing the benefit of the proposed NDFT framework for object (vehicle) detection on the UAVDT [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Our proposed NDFT-Faster-RCNN network. Algorithm 1 Learning Nuisance Disentangled Feature Transform in UAV-based Object Detection via Adversarial Training Given pre-trained NDFT module fT , object detection task module fO, and nuisances prediction modules f i N s for number of training iterations do Sample a mini-batch of n examples {X1 , · · · , Xn} Update NDFT module fT (weights wT ) and object detection module fO… view at source ↗
Figure 3
Figure 3. An example showing the benefit of the proposed NDFT approach for object detection on VisDrone2018 dataset. [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: An example showing the superior performance of NDFT-DE-FPN(r) over DE-FPN for object detection on Vis [PITH_FULL_IMAGE:figures/full_fig_p008_4.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

55 extracted references · 43 canonical work pages

  1. [1]

    https://www.kdnuggets.com/2018/09/ data-augmentation-bounding-boxes-image-transforms

    Data augmentation for bounding boxes: Re- thinking image transforms for object detec- tion. https://www.kdnuggets.com/2018/09/ data-augmentation-bounding-boxes-image-transforms. html, 2018

    work page 2018

  2. [2]

    https://www.dji.com/inspire-2/info# specs, 2018

    Dji inspire 2 specs. https://www.dji.com/inspire-2/info# specs, 2018

    work page 2018

  3. [3]

    https://scet.berkeley.edu/wp- content/uploads/ConnCarProjectReport-1.pdf, 2018

    Drones for deliveries. https://scet.berkeley.edu/wp- content/uploads/ConnCarProjectReport-1.pdf, 2018

    work page 2018

  4. [4]

    http: //aiskyeye.com/views/getInfo?loc=13, 2018

    Visdrone2018 object detection in images leaderboard. http: //aiskyeye.com/views/getInfo?loc=13, 2018

    work page 2018

  5. [5]

    Siamese-gan: Learning invariant representations for aerial vehicle image categorization

    Laila Bashmal, Yakoub Bazi, Haikel AlHichri, Mohamad M AlRahhal, Nassim Ammour, and Naif Alajlan. Siamese-gan: Learning invariant representations for aerial vehicle image categorization. Remote Sensing, 2018

    work page 2018

  6. [6]

    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), 2016

    work page 2016

  7. [7]

    Towards domain adaptive vehicle detection in satellite image by supervised super-resolution transfer

    Liujuan Cao, Rongrong Ji, Cheng Wang, and Jonathan Li. Towards domain adaptive vehicle detection in satellite image by supervised super-resolution transfer. 2016

    work page 2016

  8. [8]

    Domain adaptive faster r-cnn for object de- tection in the wild

    Yuhua Chen, Wen Li, Christos Sakaridis, Dengxin Dai, and Luc Van Gool. Domain adaptive faster r-cnn for object de- tection in the wild. In CVPR, 2018

    work page 2018

Show all 55 references
  1. [9]

    Multi-class geospatial object detection and geographic im- age classification based on collection of part detectors

    Gong Cheng, Junwei Han, Peicheng Zhou, and Lei Guo. Multi-class geospatial object detection and geographic im- age classification based on collection of part detectors. IS- PRS Journal of Photogrammetry and Remote Sensing, 2014

    2014

  2. [10]

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

    Jifeng Dai, Yi Li, Kaiming He, and Jian Sun. R-fcn: Object detection via region-based fully convolutional networks. In NeurIPS, 2016

    2016

  3. [11]

    Disentangling factors of variation via generative entangling

    Guillaume Desjardins, Aaron Courville, and Yoshua Bengio. Disentangling factors of variation via generative entangling. arXiv, 2012

    2012

  4. [12]

    The unmanned aerial vehicle benchmark: Object detection and tracking

    Dawei Du, Yuankai Qi, Hongyang Yu, Yifan Yang, Kaiwen Duan, Guorong Li, Weigang Zhang, Qingming Huang, and Qi Tian. The unmanned aerial vehicle benchmark: Object detection and tracking. arXiv, 2018

    2018

  5. [13]

    Mod- eling visual context is key to augmenting object detection datasets

    Nikita Dvornik, Julien Mairal, and Cordelia Schmid. Mod- eling visual context is key to augmenting object detection datasets. In ECCV, 2018

    2018

  6. [14]

    Uav-assisted disaster management: Applications and open issues

    Milan Erdelj and Enrico Natalizio. Uav-assisted disaster management: Applications and open issues. In Comput- ing, Networking and Communications (ICNC), 2016 Inter- national Conference on. IEEE, 2016

    2016

  7. [15]

    The pascal visual object classes (voc) challenge

    Mark Everingham, Luc Van Gool, Christopher KI Williams, John Winn, and Andrew Zisserman. The pascal visual object classes (voc) challenge. IJCV, 2010

    2010

  8. [16]

    Unsupervised domain adaptation by backpropagation

    Yaroslav Ganin and Victor Lempitsky. Unsupervised domain adaptation by backpropagation. arXiv, 2014

    2014

  9. [17]

    Fast r-cnn

    Ross Girshick. Fast r-cnn. In ICCV, 2015

    2015

  10. [18]

    Rich feature hierarchies for accurate object detection and semantic segmentation

    Ross Girshick, Jeff Donahue, Trevor Darrell, and Jitendra Malik. Rich feature hierarchies for accurate object detection and semantic segmentation. In CVPR, 2014

    2014

  11. [19]

    Image-to-image translation for cross-domain disentan- glement

    Abel Gonzalez-Garcia, Joost van de Weijer, and Yoshua Ben- gio. Image-to-image translation for cross-domain disentan- glement. arXiv preprint arXiv:1805.09730, 2018

    2018 arXiv

  12. [20]

    Generative adversarial nets

    Ian Goodfellow, Jean Pouget-Abadie, Mehdi Mirza, Bing Xu, David Warde-Farley, Sherjil Ozair, Aaron Courville, and Yoshua Bengio. Generative adversarial nets. In NeurIPS, 2014

    2014

  13. [21]

    Deep drone: object detection and tracking for smart drones on embedded system, 2016

    Song Han, William Shen, and Zuozhen Liu. Deep drone: object detection and tracking for smart drones on embedded system, 2016

    2016

  14. [22]

    Mask r-cnn

    Kaiming He, Georgia Gkioxari, Piotr Doll ´ar, and Ross Gir- shick. Mask r-cnn. In ICCV. IEEE, 2017

    2017

  15. [23]

    Spatial pyramid pooling in deep convolutional networks for visual recognition

    Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. Spatial pyramid pooling in deep convolutional networks for visual recognition. In ECCV, 2014

    2014

  16. [24]

    Cycada: Cycle-consistent adversarial domain adapta- tion

    Judy Hoffman, Eric Tzeng, Taesung Park, Jun-Yan Zhu, Phillip Isola, Kate Saenko, Alexei A Efros, and Trevor Dar- rell. Cycada: Cycle-consistent adversarial domain adapta- tion. arXiv, 2017

    2017

  17. [25]

    Processing and assessment of spectrometric, stereoscopic imagery collected using a lightweight uav spec- tral camera for precision agriculture

    Eija Honkavaara, Heikki Saari, Jere Kaivosoja, Ilkka P¨ol¨onen, Teemu Hakala, Paula Litkey, Jussi M ¨akynen, and Liisa Pesonen. Processing and assessment of spectrometric, stereoscopic imagery collected using a lightweight uav spec- tral camera for precision agriculture. Remot...

    2013

  18. [26]

    Me r- cnn: Multi-expert r-cnn for object detection

    Hyungtae Lee, Sungmin Eum, and Heesung Kwon. Me r- cnn: Multi-expert r-cnn for object detection. arXiv, 2017

    2017

  19. [27]

    Perceptual generative adversarial networks for small object detection

    Jianan Li, Xiaodan Liang, Yunchao Wei, Tingfa Xu, Jiashi Feng, and Shuicheng Yan. Perceptual generative adversarial networks for small object detection. In CVPR, 2017

    2017

  20. [28]

    Feature pyramid networks for object detection

    Tsung-Yi Lin, Piotr Doll ´ar, Ross B Girshick, Kaiming He, Bharath Hariharan, and Serge J Belongie. Feature pyramid networks for object detection

  21. [29]

    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, 2014

    2014

  22. [30]

    Enhance visual recognition under ad- verse conditions via deep networks

    Ding Liu, Bowen Cheng, Zhangyang Wang, Haichao Zhang, and Thomas S Huang. Enhance visual recognition under ad- verse conditions via deep networks. TIP, 2019

    2019

  23. [31]

    Ssd: Single shot multibox detector

    Wei Liu, Dragomir Anguelov, Dumitru Erhan, Christian Szegedy, Scott Reed, Cheng-Yang Fu, and Alexander C Berg. Ssd: Single shot multibox detector. In ECCV, 2016

    2016

  24. [32]

    Multi-task adversarial network for disentangled feature learning

    Yang Liu, Zhaowen Wang, Hailin Jin, and Ian Was- sell. Multi-task adversarial network for disentangled feature learning. In CVPR, 2018

    2018

  25. [33]

    Overview of machine learning (ml) based perception algorithms for unstructured and de- graded visual environments

    Priya Narayanan, Zhenyu Wu, Heesung Kwon, Zhangyang Wang, and Raghuveer Rao. Overview of machine learning (ml) based perception algorithms for unstructured and de- graded visual environments. In Artificial Intelligence and Machine Learning for Multi-Domain Operations Applica- t...

    2019

  26. [34]

    Sub- space alignment based domain adaptation for rcnn detector

    Anant Raj, Vinay P Namboodiri, and Tinne Tuytelaars. Sub- space alignment based domain adaptation for rcnn detector. arXiv, 2015

    2015

  27. [35]

    Vehicle detec- tion in aerial imagery: A small target detection benchmark

    S ´ebastien Razakarivony and Fr ´ed´eric Jurie. Vehicle detec- tion in aerial imagery: A small target detection benchmark. Journal of Visual Communication and Image Representation, 2016

    2016

  28. [36]

    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 Proceedings of the IEEE conference on computer vision and pattern recognition, pages 779–788, 2016

    2016

  29. [37]

    Yolo9000: better, faster, stronger

    Joseph Redmon and Ali Farhadi. Yolo9000: better, faster, stronger

  30. [38]

    Yolov3: An incremental improvement

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

    2018

  31. [39]

    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. In NeurIPS, 2015

    2015

  32. [40]

    Autonomous uav surveillance in complex urban environments

    Eduard Semsch, Michal Jakob, Du ˇsan Pavlicek, and Michal Pechoucek. Autonomous uav surveillance in complex urban environments. In Proceedings of the 2009 IEEE/WIC/ACM International Joint Conference on Web Intelligence and In- telligent Agent Technology-Volume 02. IEEE Compute...

    2009

  33. [41]

    Learning disentangled represen- tations in deep generative models

    N Siddharth, Brooks Paige, Alban Desmaison, Jan-Willem van de Meent, Frank Wood, Noah D Goodman, Pushmeet Kohli, and Philip HS Torr. Learning disentangled represen- tations in deep generative models. 2016

    2016

  34. [42]

    An analysis of scale invari- ance in object detection–snip

    Bharat Singh and Larry S Davis. An analysis of scale invari- ance in object detection–snip

  35. [43]

    Sniper: Efficient multi-scale training

    Bharat Singh, Mahyar Najibi, and Larry S Davis. Sniper: Efficient multi-scale training. arXiv, 2018

    2018

  36. [44]

    Bridging the gap between computa- tional photography and visual recognition

    Rosaura G VidalMata, Sreya Banerjee, Brandon Richard- Webster, Michael Albright, Pedro Davalos, Scott Mc- Closkey, Ben Miller, Asong Tambo, Sushobhan Ghosh, Su- darshan Nagesh, et al. Bridging the gap between computa- tional photography and visual recognition. arXiv, 2019

    2019

  37. [45]

    Privacy-preserving deep visual recog- nition: An adversarial learning framework and a new dataset

    Haotao Wang, Zhenyu Wu, Zhangyang Wang, Zhaowen Wang, and Hailin Jin. Privacy-preserving deep visual recog- nition: An adversarial learning framework and a new dataset. arXiv, 2019

    2019

  38. [46]

    A-fast-rcnn: Hard positive generation via adversary for ob- ject detection

    Xiaolong Wang, Abhinav Shrivastava, and Abhinav Gupta. A-fast-rcnn: Hard positive generation via adversary for ob- ject detection

  39. [47]

    Studying very low resolution recog- nition using deep networks

    Zhangyang Wang, Shiyu Chang, Yingzhen Yang, Ding Liu, and Thomas S Huang. Studying very low resolution recog- nition using deep networks. In CVPR, 2016

    2016

  40. [48]

    Towards privacy-preserving visual recognition via ad- versarial training: A pilot study

    Zhenyu Wu, Zhangyang Wang, Zhaowen Wang, and Hailin Jin. Towards privacy-preserving visual recognition via ad- versarial training: A pilot study. In ECCV, 2018

    2018

  41. [49]

    Dota: A large-scale dataset for object detection in aerial images

    Gui-Song Xia, Xiang Bai, Jian Ding, Zhen Zhu, Serge Be- longie, Jiebo Luo, Mihai Datcu, Marcello Pelillo, and Liang- pei Zhang. Dota: A large-scale dataset for object detection in aerial images

  42. [50]

    Linear disentangled represen- tation learning for facial actions

    Xiang Xiang and Trac D Tran. Linear disentangled represen- tation learning for facial actions. arXiv, 2017

    2017

  43. [51]

    Deep regionlets for object de- tection

    Hongyu Xu, Xutao Lv, Xiaoyu Wang, Zhou Ren, Navaneeth Bodla, and Rama Chellappa. Deep regionlets for object de- tection. In ECCV, 2018

    2018

  44. [52]

    Deep regionlets: Blended rep- resentation and deep learning for generic object detection

    Hongyu Xu, Xutao Lv, Xiaoyu Wang, Zhou Ren, Navaneeth Bodla, and Rama Chellappa. Deep regionlets: Blended rep- resentation and deep learning for generic object detection. TPAMI, 2019

    2019

  45. [53]

    Dada: Deep adversarial data augmentation for ex- tremely low data regime classification

    Xiaofeng Zhang, Zhangyang Wang, Dong Liu, and Qing Ling. Dada: Deep adversarial data augmentation for ex- tremely low data regime classification. arXiv, 2018

    2018

  46. [54]

    Vision meets drones: A challenge.arXiv, 2018

    Pengfei Zhu, Longyin Wen, Xiao Bian, Haibin Ling, and Qinghua Hu. Vision meets drones: A challenge.arXiv, 2018

    2018

  47. [55]

    Visdrone-det 2018: The vision meets drone object detection in image challenge results

    Pengfei Zhu, Longyin Wen, Dawei Du, Xiao Bian, et al. Visdrone-det 2018: The vision meets drone object detection in image challenge results. ECCV Vision Meets Drone Work- shop, 2018

    2018

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.