Pith. sign in

REVIEW 5 major objections 5 minor 52 references

Collaborative Feature-Logits Contrastive Learning for Open-Set Semi-Supervised Object Detection

T0 review · 5 major / 5 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 This paper argues that two auxiliary losses can teach a semi-supervised object detector to label out-of-distribution objects as 'unknown' rather than forcing them into known classes.

desk verdict A simple, plausible OSSOD extension of Unbiased Teacher with consistent gains, but the OOD training signal rests on an unvalidated top-K background mining mechanism and key hyperparameters are undisclosed. read the letter →

arxiv 2411.13001 v2 pith:YCZ6WPVX submitted 2024-11-20 cs.CV

Xinhao Zhong , Siyu Jiao , Yao Zhao , Yunchao Wei This is my paper · ORCID
classification cs.CV
keywords open-setsemi-supervisedobjectdetectionout-of-distributioncontrastivelearninguncertaintyclassificationlosspseudo-labelingfeaturememorypoolFasterR-CNNteacher-studenttraining
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

Open-set semi-supervised object detection means training a detector on a small labeled set of known classes while the unlabeled stream also contains objects from outside that set; standard semi-supervised detectors respond by forcing those outsiders into the closest known class. This paper argues the fix is to change the loss, not the architecture: a feature contrastive loss pulls same-class proposals together and pushes different classes apart, and an uncertainty classification loss teaches the detector to output a single 'unknown' label for out-of-distribution objects. The two losses are trained inside the ordinary teacher-student pseudo-labeling loop, so no separate out-of-distribution detector and no offline filtering stage is needed. The paper reports state-of-the-art results on COCO/VOC open-set benchmarks, including an unknown-class average precision of 9.82 on one setting, and shows the same two losses can be grafted onto another semi-supervised detector.

What carries the argument

The mechanism is a pair of auxiliary losses bolted onto the standard teacher-student pseudo-labeling loop of a proposal-based detector. $\mathcal{L}_{fc}$ is a supervised contrastive loss computed on 128-dimensional embeddings of RoI features: each proposal embedding is contrasted against a memory pool storing up to $q=256$ filtered embeddings per class, with only proposals passing $S_{IoU} > 0.7$ and $S_{cos} > 0.5$ written into the pool, and the unknown class stored without intra-class pull. $\mathcal{L}_{uc}$ is a weighted uncertainty classification loss: for known classes it is ordinary weighted cross-entropy, while for the unknown class it computes softmax over background plus unknown only and reweights the unknown term by $(1-p_k)^\alpha p_k$, where $p_k$ is the summed probability of known classes. The top-K highest background-class proposals supply the pseudo-ground-truth OOD signal during semi-supervised training. Together the two losses make the existing classifier output an 'unknown' category without changing the detector architecture.

What would settle it

On the VOC-COCO setup, count how often objects from the 60 non-VOC classes appear among the teacher's top-K background proposals versus among proposals the teacher assigns to a known class above the pseudo-label confidence threshold; if OOD objects mostly fall in the latter group, the mined signal for $\mathcal{L}_{uc}$ is empty and the reported $AP_u$ should not materialize. A second check: replace the top-K background mining with an equal number of randomly chosen background proposals; if $AP_u$ stays roughly the same, then the specific mining step is not what creates the unknown-class ability.

Watch

Extended reading notes

Core claim

The central discovery is that a closed-set semi-supervised detector can be turned into an open-set detector by joint optimization at two levels of the same model. At the feature level, the loss $\mathcal{L}_{fc}$ forms class clusters in the embedding space using a memory pool of per-class region embeddings filtered by IoU and cosine similarity; because out-of-distribution objects are semantically varied, it deliberately avoids compacting the OOD class and only separates it from known classes. At the logits level, the loss $\mathcal{L}_{uc}$ mines candidate OOD instances by taking the top-K background-class proposals and forces the classifier to assign probability mass to a unified class $K+1$ ('unknown'), with the OOD term weighted by uncertainty $(1-p_k)^\alpha p_k$. Trained in a teacher-student semi-supervised pipeline, the resulting detector reports both known classes and the unknown class; on the COCO/VOC benchmarks the paper reports $mAP_k$ up to 25.05 and $AP_u$ up to 9.82, and ablations attribute the unknown-class ability mainly to $\mathcal{L}_{uc}$.

Load-bearing premise

The method depends on the teacher detector surfacing out-of-distribution objects as high-scoring background proposals, because the uncertainty loss mines its unknown-object training signal from the top-K background boxes; if the teacher confidently assigns such objects to a known class, the loss never sees them as unknown.

Editorial extensions

If this is right

  • If the claim holds, any closed-set pseudo-label-based semi-supervised detector can be upgraded to open-set operation by adding these two losses, with no second model and no extra fine-tuning stage.
  • Detectors trained this way would stop presenting novel objects as high-confidence known classes and would instead emit a single 'unknown' label, a property directly relevant to autonomous vehicles and open-world perception systems.
  • The per-class memory pool gives the model a growing set of class prototypes during training, which should strengthen representation learning for known classes as well as separating unknowns.
  • The method remains effective with as few as 1,000 labeled images (unknown AP 4.25 on the 20/60 setting), suggesting the OOD signal from background mining is usable even when supervision is scarce.

Reading between the lines

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

  • The paper does not analyze how often OOD objects actually land in the teacher's top-K background proposals; a natural extension would be to measure that hit rate and make K adaptive, since the entire OOD training signal depends on this mine.
  • Because the losses attach to shared RoI features and logits, the same recipe should transfer to open-set versions of anchor-free and transformer-based detectors, though the paper only demonstrates transfer to one other SSOD method.
  • The acknowledged trade-off between ID accuracy and unknown accuracy implies a Pareto frontier; the paper sweeps only the loss weight $\beta$, and a systematic multi-objective study of thresholds and temperatures could find better operating points.
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 / 5 minor

Summary. The paper proposes CFL-Detector, an open-set semi-supervised object detection method built on a Faster R-CNN / Unbiased Teacher pipeline. It adds a feature contrastive loss (L_fc) that clusters embeddings of the same class and separates different classes, and an uncertainty classification loss (L_uc) that trains the detector to output a unified 'unknown' class for out-of-distribution objects. OOD supervision is obtained by mining the top-K proposals with the highest background-class probabilities in the teacher's predictions. Experiments on COCO-Open-CLS, COCO-Open-SUP, and VOC-COCO report consistent improvements in ID mAP over UT and STAC and substantially higher OOD AP_u, with ablations attributing the gains to the two proposed losses and showing transfer to STAC. The central idea is plausible and the empirical tables are internally consistent, but several load-bearing details of the training objective and the OOD-mining procedure are underspecified.

Significance. If the reported results hold, the method is a simple and practical contribution: it converts a closed-set SSOD detector into an open-set detector by mining background proposals for OOD training and adding a feature-space contrastive term, without requiring an offline OOD detector. The three benchmark settings and the transfer ablation to STAC give the central claim reasonable empirical support. The main caveat is that all OOD supervision is manufactured from top-K background mining, and the paper provides no precision/recall evaluation of that mining step; combined with missing details in the loss definitions, this makes the exact training recipe hard to reconstruct and the AP_u numbers difficult to interpret. The method is not a derivation, but the empirical claims are clear and the issues are addressable in revision.

major comments (5)
  1. [§4.2, Eq. (3)] The supervised loss in Eq. (3) is written as L_sup = L_rpn_cls + L_rpn_reg + L_roi_reg + alpha_t L_fc + beta L_uc, so it contains no ROI-head classification loss L_roi_cls. The sentence immediately below the equation says 'L_cls represents the classification loss,' but no such term appears. The uncertainty loss in Eq. (7) does include a weighted ID term w_k log(p_k), so it may be intended to replace the standard classifier loss, but then it also includes an OOD term w_u log(p_u) on labeled data where there are no OOD labels. Please specify the exact supervised objective, including how w_u is obtained on labeled data, and confirm whether the standard L_roi_cls is used or deliberately omitted. This is load-bearing because the ID mAP improvements over UT could be due to a different training recipe rather than the proposed losses.
  2. [§4.4, Eq. (7)] In the semi-supervised branch, L_uc is reduced to -sum w_u log(p_u), and L_unsup in Eq. (3) contains only RPN classification, L_fc, and L_uc. Thus, on unlabeled data the student receives no ROI-head ID classification loss and no box regression loss. This is a major deviation from the Unbiased Teacher baseline and from the standard pseudo-label pipeline. The paper should either add the usual supervised losses on pseudo-labeled proposals or provide an ablation showing that the proposed losses alone are sufficient; otherwise the reported ID mAP gains and the bounding-box quality in Fig. 4 are unexplained. If regression is intentionally disabled because pseudo-boxes are noisy, state that explicitly and quantify its effect.
  3. [§4.4, Eq. (5), Sec. 5.1] The only source of OOD supervision in the semi-supervised stage is the set of top-K proposals mined from the background class probabilities, but the value of K is never given in Sec. 5.1 or in any table, and no analysis reports how many of these K proposals are true OOD objects rather than background. If OOD objects are confidently assigned to ID classes by the teacher, they will not enter the mined set, and L_uc will teach the model to call true background 'unknown.' Please report K and provide OOD recall/precision of the mining step (e.g., what fraction of ground-truth OOD boxes appear among the top-K background proposals) on COCO-Open-CLS. Without this, the AP_u numbers in Tables 1 and 2 cannot be attributed to genuine open-set generalization.
  4. [§4.3, Eq. (4)] The feature contrastive loss L_fc requires class labels c_i for every proposal and uses S_IoU and S_cos thresholds to update the memory pool. In the unlabeled stage there are no ground-truth boxes, so it is unclear how c_i, the IoU scores, and the 'unknown' features are obtained for the memory pool. Please specify the pseudo-label generation and filtering procedure for the unsupervised contrastive branch (teacher predictions? RPN proposals with class scores? class centers from the memory pool?), and explain how OOD features are selected for storage. This is necessary to reproduce the method and to interpret the ablation in Table 4a.
  5. [§4.4, Eq. (5)] The normalization in Eq. (5) depends on c: for ID classes the denominator includes all classes, while for the OOD class it includes only background and OOD. As a result, p_u and p_k are not probabilities from the same softmax distribution, and the weighted cross-entropy terms in Eq. (7) do not correspond to a single classifier output. Please clarify the intended probabilistic interpretation or compare empirically with a single softmax over {background, ID classes, unknown}. If the asymmetric normalization is intentional, a formal justification is needed because this is the core OOD classification loss.
minor comments (5)
  1. [§3.1] 'Task Defination' should be 'Task Definition'.
  2. [§2.1] 'scherer et al.' should be 'Scheirer et al.'.
  3. [Table 5a] The row labels w_kL_ce and w_uL_ce do not map cleanly to Eq. (7); state which terms from Eq. (7) are disabled in each row and give the value of alpha used in w_u.
  4. [Figure 3] The figure is very small and the arrows/labels ('transfer', 'Select', 'Logits', 'uncertainty estimate') are hard to follow; please enlarge it and use terminology consistent with Secs. 4.3 and 4.4.
  5. [§6] The limitation paragraph says the method 'may cause a margin decline in ID class performance,' but the experiments show consistent ID improvements; please reconcile this statement with Tables 1-3 or remove it.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the losses are training objectives and the reported metrics are independent benchmark measurements.

full rationale

The paper proposes an empirical training scheme (CFL-Detector) with two loss terms, L_fc and L_uc, and evaluates them on COCO/VOC benchmarks. This is not a derivation chain in which a predicted quantity is defined in terms of an input quantity. L_uc mines top-K background proposals as pseudo-OOD training signal, but the reported AP_u is computed against held-out ground-truth OOD boxes, not against those mined proposals, so the evaluation is not equivalent to the training objective by construction. The hyperparameters (alpha_t=0.1, beta=1.0, q=256, dim=128, S_IoU=0.7, S_cos=0.5) are selected through ablations on the same benchmark, which is a standard model-selection practice and does not turn the measured mAP/AP_u into a fitted parameter renamed as a prediction. The paper does not rest on a load-bearing self-citation: related-work citations to the authors' prior papers are contextual, not used to justify the central mechanism. The undefined value of K in the background-mining step is a reproducibility concern and an empirical assumption about teacher proposals, but not a circularity: no equation in the paper reduces the final OOD detection result to the top-K mined set by construction. Overall, the central claims are self-contained empirical claims against external benchmarks, so the circularity score is 0.

Assumptions & free parameters 9 free parameters · 5 assumptions · 1 invented entities

The central claim rests on several unstated assumptions: that OOD objects show up in the background class, that one 'unknown' label captures all OOD, and that the hyperparameters chosen by ablation generalize. The method also inherits all assumptions of Unbiased Teacher and supervised contrastive learning.

free parameters (9)
  • alpha_t (L_fc weight) = 0.1 (initial, decaying)
    Chosen via ablation in Tab. 4c on the same COCO-Open-CLS setting used for final results.
  • beta (L_uc weight) = 1.0
    Chosen via ablation in Tab. 5b on the same benchmark used for final results.
  • temperature tau (contrastive) = not reported
    In Eq. 4, the temperature is a hyperparameter; the paper never gives its value.
  • memory pool size q = 256
    Chosen via ablation in Tab. 4b.
  • embedding dimension d = 128
    Chosen via ablation in Tab. 4b.
  • IoU threshold S_IoU = 0.7
    Chosen via ablation in Tab. 4d.
  • cosine threshold S_cos = 0.5
    Chosen via ablation in Tab. 4d.
  • top-K for OOD mining = unspecified
    Sec. 4.4 says 'top K candidates' but K is never defined.
  • unsupervised loss weight lambda = not reported
    Eq. 2 uses lambda without giving its value or schedule.
assumptions (5)
  • domain assumption Unbiased Teacher pseudo-labeling pipeline (teacher-student EMA with confidence threshold) is a sound basis for OSSOD.
    The method is built on UT (Sec. 3.2, 4.1) and inherits its pseudo-label quality assumptions without re-validating them in the open-set setting.
  • ad hoc to paper OOD objects are predominantly captured by the background class logits of the detector.
    Eq. 5 and Sec. 4.4 mine top-K background proposals as OOD; this is a central modeling choice specific to this paper.
  • domain assumption A single unified OOD class (K+1) adequately represents all out-of-distribution categories.
    Task definition in Sec. 3.1 sets C_u=K+1; this collapses all OOD diversity into one label.
  • domain assumption Supervised contrastive learning (SupCon) is an effective objective for separating detector embeddings by class.
    L_fc in Eq. 4 is adapted from Khosla et al. [15]; the paper relies on its known behavior.
  • ad hoc to paper A memory pool with per-class queues and fixed filtering thresholds provides reliable contrastive anchors.
    Sec. 4.3 introduces the pool and thresholds S_IoU=0.7 and S_cos=0.5; the stability of this design is not theoretically justified.
invented entities (1)
  • Unified 'unknown' class (C_u = K+1) independent evidence
    purpose: Represents all out-of-distribution objects during training and inference, giving the model a single target for OOD detection.
    It is a new label introduced by the paper's loss formulation (Sec. 3.1 and 4.4) and is evaluated via the AP_u metric on benchmarks, providing a falsifiable handle.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Collaborative Feature-Logits Contrastive Learning for Open-Set Semi-Supervised Object Detection." pith.science (2026). https://pith.science/paper/YCZ6WPVX

@misc{pith2026241113001,
  author       = {Pith},
  title        = {Pith review of: Collaborative Feature-Logits Contrastive Learning for Open-Set Semi-Supervised Object Detection},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/YCZ6WPVX}},
  note         = {Machine review of arXiv:2411.13001}
}
read the original abstract

Current Semi-Supervised Object Detection (SSOD) methods enhance detector performance by leveraging large amounts of unlabeled data, assuming that both labeled and unlabeled data share the same label space. However, in open-set scenarios, the unlabeled dataset contains both in-distribution (ID) classes and out-of-distribution (OOD) classes. Applying semi-supervised detectors in such settings can lead to misclassifying OOD class as ID classes. To alleviate this issue, we propose a simple yet effective method, termed Collaborative Feature-Logits Detector (CFL-Detector). Specifically, we introduce a feature-level clustering method using contrastive loss to clarify vector boundaries in the feature space and highlight class differences. Additionally, by optimizing the logits-level uncertainty classification loss, the model enhances its ability to effectively distinguish between ID and OOD classes. Extensive experiments demonstrate that our method achieves state-of-the-art performance compared to existing methods.

Figures

Figures reproduced from arXiv: 2411.13001 by the authors.

Figure 1
Figure 1. (a) The dataset configuration for the OSSOD task includes both labeled and unlabeled data with different distributions. [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Overview of our CFL-Detector. Training Stage 1: We begin with fully supervised pre-training on labeled data, where [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Details of the feature contrastive loss ( [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: (a) Visualization of pseudo-labels from UT (Top) and Ours (Bottom). Our method reduces OOD interference and [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

52 extracted references · 35 canonical work pages

  1. [1]

    David Berthelot, Nicholas Carlini, Ian Goodfellow, Nicolas Papernot, Avital Oliver, and Colin A Raffel. 2019. Mixmatch: A holistic approach to semi-supervised learning. Advances in neural information processing systems 32 (2019)

    2019

  2. [2]

    Kaidi Cao, Maria Brbic, and Jure Leskovec. 2021. Open-world semi-supervised learning. arXiv preprint arXiv:2102.03526 (2021)

    arXiv 2021

  3. [3]

    Mark Everingham, Luc Van Gool, Christopher KI Williams, John Winn, and Andrew Zisserman. 2010. The pascal visual object classes (voc) challenge. Inter- national journal of computer vision 88 (2010), 303–338

    2010

  4. [4]

    Yan Fang, Feng Zhu, Bowen Cheng, Luoqi Liu, Yao Zhao, and Yunchao Wei

    work page

  5. [5]

    RCNN Faster. 2015. Towards real-time object detection with region proposal networks. Advances in neural information processing systems 9199, 10.5555 (2015), 2969239–2969250

    work page 2015

  6. [6]

    Lan-Zhe Guo, Zhen-Yu Zhang, Yuan Jiang, Yu-Feng Li, and Zhi-Hua Zhou. 2020. Safe deep semi-supervised learning for unseen-class unlabeled data. In Interna- tional Conference on Machine Learning . PMLR, 3897–3906

    work page 2020

  7. [7]

    Jiaming Han, Yuqiang Ren, Jian Ding, Xingjia Pan, Ke Yan, and Gui-Song Xia

    work page

  8. [8]

    Kunyang Han, Yong Liu, Jun Hao Liew, Henghui Ding, Jiajun Liu, Yitong Wang, Yansong Tang, Yujiu Yang, Jiashi Feng, Yao Zhao, et al. 2023. Global knowledge calibration for fast open-vocabulary segmentation. InProceedings of the IEEE/CVF International Conference on Computer Vision . 797–807

    work page 2023

Show all 52 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

    2016

  2. [10]

    Jisoo Jeong, Seungeui Lee, Jeesoo Kim, and Nojun Kwak. 2019. Consistency-based semi-supervised learning for object detection. Advances in neural information processing systems 32 (2019)

    2019

  3. [11]

    Peng-Tao Jiang, Ling-Hao Han, Qibin Hou, Ming-Ming Cheng, and Yunchao Wei

  4. [12]

    Siyu Jiao, Yunchao Wei, Yaowei Wang, Yao Zhao, and Humphrey Shi. 2023. Learning mask-aware clip representations for zero-shot segmentation. Advances in Neural Information Processing Systems 36 (2023), 35631–35653

    2023

  5. [13]

    Siyu Jiao, Hongguang Zhu, Jiannan Huang, Yao Zhao, Yunchao Wei, and Humphrey Shi. 2024. Collaborative Vision-Text Representation Optimizing for Open-Vocabulary Segmentation. arXiv preprint arXiv:2408.00744 (2024)

    2024 arXiv

  6. [14]

    KJ Joseph, Salman Khan, Fahad Shahbaz Khan, and Vineeth N Balasubramanian

  7. [15]

    Prannay Khosla, Piotr Teterwak, Chen Wang, Aaron Sarna, Yonglong Tian, Phillip Isola, Aaron Maschinot, Ce Liu, and Dilip Krishnan. 2020. Supervised contrastive learning. Advances in neural information processing systems 33 (2020), 18661– 18673

    2020

  8. [16]

    Gang Li, Xiang Li, Yujie Wang, Yichao Wu, Ding Liang, and Shanshan Zhang

  9. [17]

    Hengduo Li, Zuxuan Wu, Abhinav Shrivastava, and Larry S Davis. 2022. Re- thinking pseudo labels for semi-supervised object detection. In Proceedings of the AAAI conference on artificial intelligence , Vol. 36. 1314–1322

    2022

  10. [18]

    In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition

    Towards open world object detection. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition . 5830–5840

  11. [19]

    Zekun Li, Lei Qi, Yinghuan Shi, and Yang Gao. 2023. IOMatch: Simplifying open-set semi-supervised learning with joint inliers and outliers utilization. In Proceedings of the IEEE/CVF International Conference on Computer Vision . 15870– 15879

    2023

  12. [20]

    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

    2017

  13. [21]

    In European Conference on Computer Vision

    Pseco: Pseudo labeling and consistency training for semi-supervised object detection. In European Conference on Computer Vision . Springer, 457–472

  14. [22]

    Yen-Cheng Liu, Chih-Yao Ma, Xiaoliang Dai, Junjiao Tian, Peter Vajda, Zijian He, and Zsolt Kira. 2022. Open-set semi-supervised object detection. In European Conference on Computer Vision . Springer, 143–159

    2022

  15. [23]

    Peike Li, Yunchao Wei, and Yi Yang. 2020. Meta parsing networks: Towards generalized few-shot scene parsing with adaptive metric learning. In Proceedings of the 28th ACM international conference on Multimedia . 64–72

    2020

  16. [24]

    Kuniaki Saito, Donghyun Kim, and Kate Saenko. 2021. Openmatch: Open-set semi-supervised learning with open-set consistency regularization. Advances in Neural Information Processing Systems 34 (2021), 25956–25967

    2021

  17. [25]

    Walter J Scheirer, Lalit P Jain, and Terrance E Boult. 2014. Probability models for open set recognition. IEEE transactions on pattern analysis and machine intelligence 36, 11 (2014), 2317–2324

    2014

  18. [26]

    Tsung-Yi Lin, Michael Maire, Serge Belongie, James Hays, Pietro Perona, Deva Ramanan, Piotr Dollár, and C Lawrence Zitnick. 2014. Microsoft coco: Common objects in context. In Computer Vision–ECCV 2014: 13th European Conference, Zurich, Switzerland, September 6-12, 2014, Proce...

    2014

  19. [27]

    Kihyuk Sohn, Zizhao Zhang, Chun-Liang Li, Han Zhang, Chen-Yu Lee, and Tomas Pfister. 2020. A simple semi-supervised learning framework for object detection. arXiv preprint arXiv:2005.04757 (2020)

    2020 arXiv

  20. [28]

    Yen-Cheng Liu, Chih-Yao Ma, Zijian He, Chia-Wen Kuo, Kan Chen, Peizhao Zhang, Bichen Wu, Zsolt Kira, and Peter Vajda. 2021. Unbiased teacher for semi-supervised object detection. arXiv preprint arXiv:2102.09480 (2021)

    2021 arXiv

  21. [29]

    Yihe Tang, Weifeng Chen, Yijun Luo, and Yuting Zhang. 2021. Humble teachers teach better students for semi-supervised object detection. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition . 3132–3141

    2021

  22. [30]

    Yuxing Tang, Josiah Wang, Boyang Gao, Emmanuel Dellandréa, Robert Gaizauskas, and Liming Chen. 2016. Large scale semi-supervised object de- tection using visual and semantic knowledge transfer. In Proceedings of the IEEE conference on computer vision and pattern recognition . ...

    2016

  23. [31]

    Kihyuk Sohn, David Berthelot, Nicholas Carlini, Zizhao Zhang, Han Zhang, Colin A Raffel, Ekin Dogus Cubuk, Alexey Kurakin, and Chun-Liang Li. 2020. Fixmatch: Simplifying semi-supervised learning with consistency and confidence. Advances in neural information processing systems...

    2020

  24. [32]

    Zhenyu Wang, Yali Li, Ye Guo, Lu Fang, and Shengjin Wang. 2021. Data- uncertainty guided multi-phase learning for semi-supervised object detection. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recogni- tion. 4568–4577

    2021

  25. [33]

    Peng Tang, Chetan Ramaiah, Yan Wang, Ran Xu, and Caiming Xiong. 2021. Proposal learning for semi-supervised object detection. In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision . 2291–2301

    2021

  26. [34]

    Qizhe Xie, Minh-Thang Luong, Eduard Hovy, and Quoc V Le. 2020. Self- training with noisy student improves imagenet classification. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition . 10687–10698

    2020

  27. [35]

    Qize Yang, Xihan Wei, Biao Wang, Xian-Sheng Hua, and Lei Zhang. 2021. Inter- active self-training with mean teachers for semi-supervised object detection. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition . 5941–5950

    2021

  28. [36]

    Antti Tarvainen and Harri Valpola. 2017. Mean teachers are better role models: Weight-averaged consistency targets improve semi-supervised deep learning results. Advances in neural information processing systems 30 (2017)

    2017

  29. [37]

    Gengwei Zhang, Guoliang Kang, Yi Yang, and Yunchao Wei. 2021. Few-shot segmentation via cycle-consistent transformer. Advances in Neural Information Processing Systems 34 (2021), 21984–21996

    2021

  30. [38]

    Zerun Wang, Ling Xiao, Liuyu Xiang, Zhaotian Weng, and Toshihiko Yamasaki

  31. [39]

    arXiv preprint arXiv:2305.13802 (2023)

    Online Open-set Semi-supervised Object Detection via Semi-supervised Outlier Filtering. arXiv preprint arXiv:2305.13802 (2023)

    2023 arXiv

  32. [40]

    Yabo Zhang, Yuxiang Wei, Dongsheng Jiang, Xiaopeng Zhang, Wangmeng Zuo, and Qi Tian. 2023. Controlvideo: Training-free controllable text-to-video genera- tion. arXiv preprint arXiv:2305.13077 (2023)

    2023 arXiv

  33. [41]

    Mingkai Zheng, Shan You, Lang Huang, Fei Wang, Chen Qian, and Chang Xu

  34. [42]

    Bingfeng Zhang, Jimin Xiao, Yunchao Wei, and Yao Zhao. 2023. Credible dual- expert learning for weakly supervised semantic segmentation. International Journal of Computer Vision 131, 8 (2023), 1892–1908

    2023

  35. [43]

    Hongguang Zhu, Yunchao Wei, Xiaodan Liang, Chunjie Zhang, and Yao Zhao

  36. [44]

    Gengwei Zhang, Shant Navasardyan, Ling Chen, Yao Zhao, Yunchao Wei, Humphrey Shi, et al. 2022. Mask matching transformer for few-shot segmentation. Advances in Neural Information Processing Systems 35 (2022), 823–836

    2022

  37. [45]

    Gengwei Zhang, Liyuan Wang, Guoliang Kang, Ling Chen, and Yunchao Wei

  38. [46]

    In Proceedings of the IEEE/CVF International Conference on Computer Vision

    Slca: Slow learner with classifier alignment for continual learning on a pre-trained model. In Proceedings of the IEEE/CVF International Conference on Computer Vision. 19148–19158

  39. [49]

    In Pro- ceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition

    Simmatch: Semi-supervised learning with similarity matching. In Pro- ceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition . 14471–14481

  40. [50]

    Qiang Zhou, Chaohui Yu, Zhibin Wang, Qi Qian, and Hao Li. 2021. Instant- teaching: An end-to-end semi-supervised object detection framework. In Pro- ceedings of the IEEE/CVF conference on computer vision and pattern recognition . 4081–4090

    2021

  41. [52]

    In Proceedings of the IEEE/CVF International Conference on Computer Vision

    Ctp: Towards vision-language continual pretraining via compatible mo- mentum contrast and topology preservation. In Proceedings of the IEEE/CVF International Conference on Computer Vision . 22257–22267

  42. [2021]

    IEEE Transactions on Pattern Analysis and Machine Intelligence 44, 10 (2021), 7062–7077

    Online attention accumulation for weakly supervised semantic segmenta- tion. IEEE Transactions on Pattern Analysis and Machine Intelligence 44, 10 (2021), 7062–7077

    2021

  43. [2022]

    In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition

    Expanding low-density latent regions for open-set object detection. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 9591–9600

  44. [2023]

    In Proceedings of the IEEE/CVF International Conference on Computer Vision

    Locating noise is halfway denoising for semi-supervised segmentation. In Proceedings of the IEEE/CVF International Conference on Computer Vision . 16612– 16622

Pith tools

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