REVIEW 4 major objections 6 minor 40 references
Self-Training and Adversarial Background Regularization for Unsupervised Domain Adaptive One-Stage Object Detection
T0 review · 4 major / 6 minor · reviewed 2026-08-14 · deepseek-v4-flash
Pith's one-line read A two-part recipe adapts one-stage object detectors to new domains with no target labels.
desk verdict A useful empirical recipe for one-stage UDA detection, but the headline Clipart1k gain is transductive and the paper needs a cleaner unseen-target evaluation before the numbers are trusted. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The load-bearing machinery is the interaction between two losses on target images. WST replaces the standard detection loss with $L_{ST}(x_t,\hat{c}) = -\sum_{i \in Pos}\log(p_i(\hat{c}_i|x_t))-\sum_{i \in \tilde{Neg}}\log(p_i(0|x_t))$, where Pos are examples selected by SRRS, a region-agreement score that averages the confidence of all boxes with IoU greater than a threshold to a final detection, and $\tilde{Neg}$ are the easiest negatives from the mined negative set. BSR adds an adversarial background loss with a focal term, $L_{adv}(x_t) = -t\sum_i |t-p_i(0|x_t)|^\gamma \log(p_i(0|x_t)) - (1-t)\sum_i |t-p_i(0|x_t)|^\gamma \log(1-p_i(0|x_t))$, applied through a gradient reversal layer after relu4_3 of SSD300 to the target examples with the lowest background scores. The first mechanism stabilizes self-training by filtering unreliable pseudo-labels; the second forces the feature extractor to produce target background features that the classifier cannot confidently label as background, thereby sharpening foreground-background separation.
What would settle it
Run the adapted model on a clipart test set disjoint from the images used in WST and BSR training; if the mAP returns to near the 26.7 source-only baseline, the Clipart1k result came from fitting the evaluation images rather than from domain adaptation itself. Watercolor2k and Comic2k already provide held-out test sets, so their reported gains can be checked directly against unseen images.
Extended reading notes
Core claim
The paper's central claim is that one-stage object detectors can be adapted to a new visual domain without any target-domain labels by combining two complementary regularizers: weak self-training (WST) and adversarial background score regularization (BSR). WST generates pseudo-labels only for detections whose Supporting Region-based Reliable Score (SRRS) exceeds a threshold, omits regression updates on unreliable boxes, and replaces hard negative mining with weak negative mining so that confidently background-like examples are used as negatives. BSR treats the background probability output as an adversarial game: the classifier is pushed toward predicting $t=0.5$ for selected target detections, while the feature extractor tries to make background predictions confident, with a focal term controlling the strength. Starting from SSD300 trained on Pascal VOC, the combined method improves mAP from 26.7 to 35.7 on Clipart1k, from 47.1 to 49.9 on Watercolor2k, and from 21.9 to 26.8 on Comic2k, whereas naive self-training collapses to 10.7 mAP on Clipart1k.
Load-bearing premise
The reported gains on Clipart1k are measured on the same images used for adaptation, so the method may not generalize to unseen target images.
Editorial extensions
If this is right
- On all three target sets, the proposed WST and BSR combination beats both the source-only SSD baseline and a domain-adversarial feature alignment baseline, so one-stage detectors can benefit from target-domain training without any target labels.
- Naive pseudo-label training on target images can be worse than no adaptation at all (10.7 mAP on Clipart1k versus a 26.7 baseline), so reliability filtering and conservative negative sampling are what make self-training usable for detection.
- Background-score regularization alone already improves over the baseline and over the domain-adversarial baseline on Clipart1k (34.0 mAP), suggesting that foreground-background separation is a productive place to attack domain shift in one-stage detectors.
- The two components are complementary: BSR improves discrimination between foreground and background, while WST supplies class information, so combining them yields the largest gains rather than either alone.
Reading between the lines
- Because the Clipart1k evaluation uses the same images for adaptation and testing, the 9-point mAP gain there may partly reflect fitting those particular images; a held-out clipart split would give a cleaner estimate of how the method transfers to unseen target images.
- WST's weak negative mining could be combined with a curriculum that starts with very conservative negatives and relaxes as pseudo-labels stabilize; the paper does not explore this schedule.
- BSR needs no separate discriminator network, only a gradient reversal layer on the detector's own background output, so the same idea could be applied to other dense one-stage detectors beyond SSD.
Signed reviews
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. This paper addresses unsupervised domain adaptation for one-stage object detection. The authors build on SSD300 and propose two components: weak self-training (WST) and adversarial background score regularization (BSR). WST generates pseudo-labels using a Supporting Region-based Reliable Score (SRRS), masks gradients of hard negatives, and selects only a subset of easy negatives for the background loss, while omitting the localization loss for pseudo-labels. BSR adds a gradient-reversal layer after relu4_3 and an adversarial loss on background probabilities with a focal term, intended to make target features more discriminative. Experiments adapt from Pascal VOC to Clipart1k, Watercolor2k, and Comic2k, reporting mAP gains over the SSD baseline (e.g., 26.7 to 35.7 on Clipart1k, 47.1 to 49.9 on Watercolor2k, 21.9 to 26.8 on Comic2k). Ablations on WST and parameter sensitivity on BSR are also reported.
Significance. If the reported gains are taken at face value, the paper provides a useful empirical recipe for pseudo-labeling and adversarial background regularization in one-stage UDA detection, and the ablation study (Table 4) supports the claim that controlling both false positives and false negatives is important. The components are simple and do not require additional networks beyond a gradient reversal layer, which is a practical advantage. However, the strongest evidence, the Clipart1k result, is obtained under a transductive protocol in which the same images are used for adaptation and evaluation, and several hyperparameters and the early-stopping point are selected with target performance in view. These issues must be resolved before the central claim about generalization to the target domain is established.
major comments (4)
- [Section 4.1, Table 1] The Clipart1k evaluation is transductive: the manuscript states that "We used all images as a target dataset both for training and evaluation." The largest reported improvement (26.7 to 35.7 mAP, +9.0) is therefore measured on the same images used for adaptation, and it does not demonstrate that the detector improves on unseen target-domain images. Watercolor2k and Comic2k use separate train/test splits, so the issue is specific to Clipart1k, but the headline gain is on Clipart1k. Please either evaluate on a held-out target split or explicitly reframe the contribution and claims as transductive UDA; in either case, the abstract and conclusion should not claim unsupervised domain adaptation performance on unseen target data based on this protocol.
- [Section 4.2, BSR with WST] The early-stopping rule is not independent of the test set. The text says training was "early stopped at 55k iterations since self-training is not helpful when it is overused," but no validation split is described. Because Clipart1k is used both for training and evaluation, the stopping iteration is effectively chosen by looking at the target test curve, which inflates the reported mAP and the ablation comparisons in Section 5. Please report a validation-based stopping rule or a fixed schedule, and state which number is used in each table.
- [Section 4.3, Tables 2–3] Several dataset-specific hyperparameters appear to be selected using target test performance: the Watercolor2k self-training learning rate is set to 1e-6 because of that dataset's properties, and gamma=3.0 is used only for Comic2k. Without a validation split or a sensitivity analysis for these choices, the cross-dataset consistency of the method is not established. Please fix hyperparameters across datasets or justify them with validation-set experiments.
- [Section 5, Tables 4–5] All ablation and sensitivity experiments are conducted only on Clipart1k, which has the transductive protocol described above. Consequently, the conclusions that both SRRS and weak negative mining are essential, and that t=0.5 is the best BSR target, are based on the same images used for adaptation and early stopping. These conclusions need to be re-validated on a held-out target split or at least on Watercolor2k/Comic2k, where separate train/test sets exist.
minor comments (6)
- [Section 4.2] The schedule "epsilon = 1/(1+e^{-3p})" uses "maxiteration", but Section 4.2 does not define maxiteration for the BSR+WST run; if it is the early-stopping point (55k), the epsilon schedule and the stopping rule are entangled and should be clarified.
- [Figure 2] The curves in Figure 2 are difficult to read; add axis labels, a legend, and larger line styles so the claim about naive self-training degenerating can be checked.
- [Tables 1–3] The checkmark rows do not identify which component is active; please label rows as "Ours (BSR)", "Ours (WST)", and "Ours (BSR+WST)" in each table.
- [Section 3.3] The name "background score regularization" is confusing because the selected examples have the lowest background scores (ascending order), i.e., they are the most foreground-like detections; clarify in the text why these examples are the right ones for regularizing background discrimination.
- [Section 5.2] The parameter sensitivity table does not state which hyperparameters are held fixed during the t-sweep (e.g., gamma=2.0, same early stopping), so the comparison is not fully reproducible; state the fixed settings explicitly.
- [Section 2.3] The claim that PL [15] "is not valid" in the unsupervised setting is stated without experimental evidence; report the actual numbers or omit the claim.
Circularity Check
No circular derivation: the WST and BSR mechanisms are empirically defined and ablated, with no equation-level reduction to their inputs; the transductive Clipart1k evaluation is a validity caveat, not circularity.
full rationale
The paper makes an empirical claim and does not derive its improvements from definitions or from self-citations. Eqs. (2) and (3) define WST: a masked self-training loss and an IoU-weighted reliability score; these are stated mechanisms, not restatements of the reported mAP. Eq. (7) defines BSR as an adversarial background-score loss; the reported gains are ablated in Tables 1-4, so the components have independent content. Self-citations [5] and [16] appear only in related-work lists or as prior detection work and are not load-bearing for the WST/BSR design. The strongest caveat is Section 4.1, where for Clipart1k 'we used all images as a target dataset both for training and evaluation,' and Section 4.2, where training is 'early stopped at 55k iterations' without a described held-out validation split; this weakens the external validity of the headline +9.0 mAP result but is an evaluation-protocol concern, not a circular derivation. Therefore there is no significant circularity to report.
Assumptions & free parameters
free parameters (8)
- SRRS threshold epsilon =
0.8 for WST; 1/(1+e^{-3p}) schedule when combined with BSR
- IoU support threshold delta =
0.5
- BSR target background probability t =
0.5
- BSR focal exponent gamma =
2.0 on Clipart1k and Watercolor2k, 3.0 on Comic2k
- Weak negative mining ratio =
|Neg|/3
- Early stopping iteration for BSR+WST =
55k
- WST learning rate =
1e-5; 1e-6 for Watercolor2k
- BSR negative sampling ratio 3N =
lowest background-score 3N examples per batch
assumptions (4)
- domain assumption Source and target share the same label space and feature representation, and pseudo-labels with high SRRS are reliable enough to train on.
- domain assumption Backgrounds of source and target share less common features than foregrounds, so pushing target background scores toward 0.5 is a useful regularizer.
- domain assumption Hard negative mining on target pseudo-labels mostly selects false negatives rather than true backgrounds.
- standard math Gradient reversal layer implements the min-max objective in Eqs. (5)-(6).
Cite this review
Pith. "Pith review of Self-Training and Adversarial Background Regularization for Unsupervised Domain Adaptive One-Stage Object Detection." pith.science (2026). https://pith.science/paper/43TPJVVD
@misc{pith2026190900597,
author = {Pith},
title = {Pith review of: Self-Training and Adversarial Background Regularization for Unsupervised Domain Adaptive One-Stage Object Detection},
year = {2026},
howpublished = {\url{https://pith.science/paper/43TPJVVD}},
note = {Machine review of arXiv:1909.00597}
}
read the original abstract
Deep learning-based object detectors have shown remarkable improvements. However, supervised learning-based methods perform poorly when the train data and the test data have different distributions. To address the issue, domain adaptation transfers knowledge from the label-sufficient domain (source domain) to the label-scarce domain (target domain). Self-training is one of the powerful ways to achieve domain adaptation since it helps class-wise domain adaptation. Unfortunately, a naive approach that utilizes pseudo-labels as ground-truth degenerates the performance due to incorrect pseudo-labels. In this paper, we introduce a weak self-training (WST) method and adversarial background score regularization (BSR) for domain adaptive one-stage object detection. WST diminishes the adverse effects of inaccurate pseudo-labels to stabilize the learning procedure. BSR helps the network extract discriminative features for target backgrounds to reduce the domain shift. Two components are complementary to each other as BSR enhances discrimination between foregrounds and backgrounds, whereas WST strengthen class-wise discrimination. Experimental results show that our approach effectively improves the performance of the one-stage object detection in unsupervised domain adaptation setting.
Figures
Figures from the paper (3 more)
Reference graph
Works this paper leans on
-
[1]
Semi-supervised learning (chapelle, o
Olivier Chapelle, Bernhard Scholkopf, and Alexander Zien. Semi-supervised learning (chapelle, o. et al., eds.; 2006)[book reviews]. IEEE Transactions on Neural Net- works, 20(3):542–542, 2009
work page 2006
-
[2]
Co- training for domain adaptation
Minmin Chen, Kilian Q Weinberger, and John Blitzer. Co- training for domain adaptation. In Advances in neural infor- mation processing systems, pages 2456–2464, 2011
work page 2011
-
[3]
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. InProceedings of the IEEE conference on computer vision and pattern recognition, pages 3339–3348, 2018
work page 2018
-
[4]
Adaptive Semantic Segmentation with a Strategic Curriculum of Proxy Labels
Kashyap Chitta, Jianwei Feng, and Martial Hebert. Adaptive semantic segmentation with a strategic curriculum of proxy labels. arXiv preprint arXiv:1811.03542, 2018
work page Pith review arXiv 2018
-
[5]
Pseudo-labeling curriculum for unsupervised domain adaptation
Jaehoon Choi, Minki Jeong, Taekyung Kim, and Changick Kim. Pseudo-labeling curriculum for unsupervised domain adaptation. arXiv preprint arXiv:1908.00262, 2019
arXiv 1908
-
[6]
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 Advances in neural information processing systems , pages 379–387, 2016
2016
-
[7]
Imagenet: A large-scale hierarchical image database
Jia Deng, Wei Dong, Richard Socher, Li-Jia Li, Kai Li, and Li Fei-Fei. Imagenet: A large-scale hierarchical image database. In 2009 IEEE conference on computer vision and pattern recognition, pages 248–255. Ieee, 2009
2009
-
[8]
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. International journal of computer vision, 88(2):303–338, 2010
2010
Show all 40 references
-
[9]
Self-ensembling for visual domain adaptation
Geoffrey French, Michal Mackiewicz, and Mark Fisher. Self-ensembling for visual domain adaptation. arXiv preprint arXiv:1706.05208, 2017
2017 arXiv
-
[10]
Domain-adversarial train- ing of neural networks
Yaroslav Ganin, Evgeniya Ustinova, Hana Ajakan, Pas- cal Germain, Hugo Larochelle, Franc ¸ois Laviolette, Mario Marchand, and Victor Lempitsky. Domain-adversarial train- ing of neural networks. The Journal of Machine Learning Research, 17(1):2096–2030, 2016
2016
-
[11]
Fast r-cnn
Ross Girshick. Fast r-cnn. In Proceedings of the IEEE inter- national conference on computer vision , pages 1440–1448, 2015
2015
-
[12]
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 Proceedings of the IEEE con- ference on computer vision and pattern recognition , pages 580–587, 2014
2014
-
[13]
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 preprint arXiv:1711.03213, 2017
2017 arXiv
-
[14]
Conditional generative adversarial network for struc- tured domain adaptation
Weixiang Hong, Zhenzhen Wang, Ming Yang, and Junsong Yuan. Conditional generative adversarial network for struc- tured domain adaptation. In Proceedings of the IEEE Con- ference on Computer Vision and Pattern Recognition, pages 1335–1344, 2018
2018
-
[15]
Cross-domain weakly-supervised object de- tection through progressive domain adaptation
Naoto Inoue, Ryosuke Furuta, Toshihiko Yamasaki, and Kiy- oharu Aizawa. Cross-domain weakly-supervised object de- tection through progressive domain adaptation. In Proceed- ings of the IEEE conference on computer vision and pattern recognition, pages 5001–5009, 2018
2018
-
[16]
Diversify and match: A domain adaptive representation learning paradigm for object detec- tion
Taekyung Kim, Minki Jeong, Seunghyeon Kim, Seokeon Choi, and Changick Kim. Diversify and match: A domain adaptive representation learning paradigm for object detec- tion. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pages 12456–12465, 2019
2019
-
[17]
Pseudo-label: The simple and efficient semi-supervised learning method for deep neural networks
Dong-Hyun Lee. Pseudo-label: The simple and efficient semi-supervised learning method for deep neural networks. In Workshop on Challenges in Representation Learning, ICML, volume 3, page 2, 2013
2013
-
[18]
Focal loss for dense object detection
Tsung-Yi Lin, Priya Goyal, Ross Girshick, Kaiming He, and Piotr Doll´ar. Focal loss for dense object detection. In Pro- ceedings of the IEEE international conference on computer vision, pages 2980–2988, 2017
2017
-
[19]
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 European conference on computer vision , pages 740–755. Springer, 2014
2014
-
[20]
Receptive field block net for accurate and fast object detection
Songtao Liu, Di Huang, et al. Receptive field block net for accurate and fast object detection. In Proceedings of the Eu- ropean Conference on Computer Vision (ECCV), pages 385– 400, 2018
2018
-
[21]
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 European con- ference on computer vision, pages 21–37. Springer, 2016
2016
-
[22]
Learning transferable features with deep adaptation networks
Mingsheng Long, Yue Cao, Jianmin Wang, and Michael I Jordan. Learning transferable features with deep adaptation networks. arXiv preprint arXiv:1502.02791, 2015
2015 arXiv
-
[23]
Conditional adversarial domain adapta- tion
Mingsheng Long, Zhangjie Cao, Jianmin Wang, and Michael I Jordan. Conditional adversarial domain adapta- tion. In Advances in Neural Information Processing Systems, pages 1640–1650, 2018
2018
-
[24]
Transfer feature learning with joint distribution adaptation
Mingsheng Long, Jianmin Wang, Guiguang Ding, Jiaguang Sun, and Philip S Yu. Transfer feature learning with joint distribution adaptation. In Proceedings of the IEEE inter- national conference on computer vision , pages 2200–2207, 2013
2013
-
[25]
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
-
[26]
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 Advances in neural information pro- cessing systems, pages 91–99, 2015
2015
-
[27]
Asymmetric tri-training for unsupervised domain adaptation
Kuniaki Saito, Yoshitaka Ushiku, and Tatsuya Harada. Asymmetric tri-training for unsupervised domain adaptation. In Proceedings of the 34th International Conference on Ma- chine Learning-Volume 70 , pages 2988–2997. JMLR. org, 2017
2017
-
[28]
Strong-weak distribution alignment for adaptive object detection
Kuniaki Saito, Yoshitaka Ushiku, Tatsuya Harada, and Kate Saenko. Strong-weak distribution alignment for adaptive object detection. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition , pages 6956– 6965, 2019
2019
-
[29]
Open set domain adaptation by backpropa- gation
Kuniaki Saito, Shohei Yamamoto, Yoshitaka Ushiku, and Tatsuya Harada. Open set domain adaptation by backpropa- gation. In Proceedings of the European Conference on Com- puter Vision (ECCV), pages 153–168, 2018
2018
-
[30]
Self paced deep learning for weakly supervised ob- ject detection
Enver Sangineto, Moin Nabi, Dubravko Culibrk, and Nicu Sebe. Self paced deep learning for weakly supervised ob- ject detection. IEEE transactions on pattern analysis and machine intelligence, 41(3):712–725, 2018
2018
-
[31]
Learning from synthetic data: Addressing domain shift for semantic segmentation
Swami Sankaranarayanan, Yogesh Balaji, Arpit Jain, Ser Nam Lim, and Rama Chellappa. Learning from synthetic data: Addressing domain shift for semantic segmentation. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pages 3752–3761, 2018
2018
-
[32]
A dirt-t approach to unsupervised domain adaptation
Rui Shu, Hung H Bui, Hirokazu Narui, and Stefano Ermon. A dirt-t approach to unsupervised domain adaptation. arXiv preprint arXiv:1802.08735, 2018
2018 arXiv
-
[33]
Adversarial discriminative domain adaptation
Eric Tzeng, Judy Hoffman, Kate Saenko, and Trevor Dar- rell. Adversarial discriminative domain adaptation. In Pro- ceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pages 7167–7176, 2017
2017
-
[34]
Learning semantic representations for unsupervised domain adaptation
Shaoan Xie, Zibin Zheng, Liang Chen, and Chuan Chen. Learning semantic representations for unsupervised domain adaptation. In International Conference on Machine Learn- ing, pages 5419–5428, 2018
2018
-
[35]
Single-shot refinement neural network for object detection
Shifeng Zhang, Longyin Wen, Xiao Bian, Zhen Lei, and Stan Z Li. Single-shot refinement neural network for object detection. In Proceedings of the IEEE Conference on Com- puter Vision and Pattern Recognition , pages 4203–4212, 2018
2018
-
[36]
Collaborative and adversarial network for unsupervised do- main adaptation
Weichen Zhang, Wanli Ouyang, Wen Li, and Dong Xu. Collaborative and adversarial network for unsupervised do- main adaptation. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition , pages 3801– 3809, 2018
2018
-
[37]
Zigzag learning for weakly supervised object detection
Xiaopeng Zhang, Jiashi Feng, Hongkai Xiong, and Qi Tian. Zigzag learning for weakly supervised object detection. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pages 4262–4270, 2018
2018
-
[38]
Weakly-supervised object detec- tion via mining pseudo ground truth bounding-boxes
Yongqiang Zhang, Yaicheng Bai, Mingli Ding, Yongqiang Li, and Bernard Ghanem. Weakly-supervised object detec- tion via mining pseudo ground truth bounding-boxes. Pat- tern Recognition, 84:68–81, 2018
2018
-
[39]
Scale-transferrable object detection
Peng Zhou, Bingbing Ni, Cong Geng, Jianguo Hu, and Yi Xu. Scale-transferrable object detection. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pages 528–537, 2018
2018
-
[40]
Unsupervised domain adaptation for semantic seg- mentation via class-balanced self-training
Yang Zou, Zhiding Yu, BVK Vijaya Kumar, and Jinsong Wang. Unsupervised domain adaptation for semantic seg- mentation via class-balanced self-training. In Proceedings of the European Conference on Computer Vision (ECCV) , pages 289–305, 2018
2018
Reviewed August 14, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.