Pith. sign in

REVIEW 4 major objections 5 minor 43 references

Unsupervised Domain Adaptive Person Search via Dual Self-Calibration

T0 review · 4 major / 5 minor · reviewed 2026-08-11 · deepseek-v4-flash

Pith's one-line read Cleaning pseudo-label noise at both image and instance levels pushes unsupervised domain-adaptive person search to state-of-the-art, rivaling some fully supervised models.

desk verdict Solid incremental UDA person search with useful ablations; the filter mechanism is plausible but not directly proven, and single-seed results keep the SOTA claim modest. read the letter →

arxiv 2412.16506 v1 pith:SYKDAT5V submitted 2024-12-21 cs.CV

classification cs.CV
keywords unsuperviseddomainadaptationpersonsearchpseudo-labeldenoisingadaptivefeaturefilteringclusterproxymemorycontrastivelearningpedestriandetectionre-identification
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

This paper tries to establish that unsupervised domain-adaptive person search can be made substantially more accurate by treating the two sources of pseudo-label noise separately: low-quality pseudo bounding boxes from the detector and misassigned pseudo identities from clustering. The proposed framework filters noisy boxes out of image-level features with an adaptive threshold in the backbone, and it cleans instance-level noise by replacing per-instance memory entries with one proxy representation per cluster. On the CUHK-SYSU and PRW benchmarks the authors report state-of-the-art results among UDA methods, with margins that make the approach competitive with some fully and weakly supervised systems. The reason this matters is that a UDA person search model that no longer drowns in its own noisy labels is closer to being usable on new camera networks without manual annotation.

What carries the argument

Two mechanisms carry the argument. First, the Perception-Driven Threshold derives a dynamic channel-wise cutoff from average pooling scaled by a sigmoid of an MLP on max pooling, and the Self-Calibrating Filter applies this cutoff through Eq. 6, a learnable blend of the classical soft threshold and a high-order soft threshold that approaches $(x^2 - \tau^2)^{1/2}$. This filter block is inserted after each of the first four ResNet-50 layers to suppress background and noisy pseudo-box features. Second, the Cluster Proxy Representation keeps one proxy $c_k$ per identity cluster in a memory dictionary, initializes each proxy as the centroid of its cluster, updates it online by $c_k \leftarrow (1-\gamma)c_k + \gamma f$, and re-clusters once per epoch with an EMA-matched offline reinitialization. The ablations show that removing either mechanism lowers mAP and top-1 on both datasets, which is why the paper attributes the gain to these two components.

What would settle it

Run the trained model on the target-domain test set and, for every ground-truth person box, compute the fraction of feature channels that the Self-Calibrating Filter sets to zero; if a substantial fraction of channels on correctly detected people are zeroed out, the filter is removing foreground signal rather than only noise, contradicting the paper's explanation of its mechanism.

Watch

Extended reading notes

Core claim

The central claim, stated on the paper's own terms, is that noisy pseudo-labels in unsupervised domain-adaptive person search are best handled by a dual self-calibration: filter unreliable image-level features in the backbone, and de-pollute the cluster memory used for identity learning. The Perception-Driven Adaptive Filter computes a per-channel threshold $\tau = \alpha \odot \operatorname{AvgPool}(F)$ with $\alpha = \sigma(\operatorname{MLP}(\operatorname{MaxPool}(F)))$, then applies a self-calibrating high-order soft threshold that interpolates from a soft to a hard filter over training. The Cluster Proxy Representation stores a single momentum-updated proxy per cluster instead of a memory slot per instance, clusters the target domain only once per epoch, and reinitializes proxies offline with exponential moving average smoothing. According to the reported experiments, these modules together reach 80.2% mAP and 81.7% top-1 on CUHK-SYSU and 39.9% mAP and 81.6% top-1 on PRW, surpassing prior UDA methods and matching or beating several weakly and fully supervised baselines.

Load-bearing premise

The load-bearing premise is that a channel's average activation, scaled by a factor learned from its max activation, separates foreground person features from background and noisy-box features in every backbone layer and across both domains; if that separation fails after domain shift, the filter would discard useful identity information while the end-to-end ablations would not reveal where the loss occurs.

Editorial extensions

If this is right

  • If noisy pseudo-labels are the main bottleneck, further UDA gains should come from better denoising of boxes and identities rather than from more elaborate domain alignment alone.
  • Because the memory dictionary holds one proxy per cluster rather than one entry per instance, training cost and memory usage scale with the number of identities, not the number of detected boxes.
  • The filter is inserted inside the backbone, so the same PDAF module can be added to other end-to-end person-search architectures without changing detection or ReID heads.
  • Reaching parity with weakly supervised and some fully supervised baselines on these benchmarks means annotation-free person search may be practical for camera networks with modest domain shift.

Reading between the lines

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

  • One could test whether the Perception-Driven Threshold transfers to other detection-and-retrieval tasks, such as vehicle search, since the threshold depends only on channel statistics rather than person-specific priors.
  • A natural extension is to make the cluster proxy update fully online, removing the once-per-epoch re-clustering step; the risk is that drifting proxies would reintroduce the misidentification pollution CPR is designed to erase.
  • A stress test that artificially corrupts a known fraction of pseudo-boxes, holding everything else fixed, would isolate how much of PDAF's benefit comes from filtering exactly this type of noise.
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 manuscript proposes Dual Self-Calibration (DSCA), an unsupervised domain-adaptive person search framework built on a DAPS-style end-to-end detector. It introduces (i) a Perception-Driven Adaptive Filter (PDAF), comprising a Perception-Driven Threshold (PDT) in Eq. (2) and a Self-Calibrating Filter (SCF) in Eq. (6), which suppresses backbone features below a channel-wise adaptive threshold; and (ii) a Cluster Proxy Representation (CPR), which replaces instance-level memory with cluster-level proxies using the updates in Eqs. (8)-(10). Experiments on CUHK-SYSU and PRW report 80.2% mAP / 81.7% top-1 and 39.9% mAP / 81.6% top-1, outperforming DAPS and DDAM and also exceeding several weakly supervised baselines. Ablations in Tables 2-3 and visualizations in Fig. 7 are used to argue that each component contributes to the final result.

Significance. The significance is moderate-to-high if the reported results hold: the method advances the UDA person search state of the art, especially on PRW, while simplifying training through one clustering per epoch. The code release and the end-to-end ablations for both modules on both benchmarks are strengths, and the paper makes no hidden circular assumptions: the modules and ablations are defined independently of the reported numbers. However, the empirical case is limited by single-seed runs, per-dataset epoch schedules, and the absence of direct evidence for the proposed filtering mechanism, so the central claims should be treated as promising rather than fully established.

major comments (4)
  1. [§3.2, Eqs. (2) and (6)] The central causal claim that PDAF removes noisy pseudo-boxes rather than distorting true foreground features is not directly supported. The threshold τ = α ⊙ AvgPool(F) is a scaled channel mean; in a target frame with a large background area this threshold can exceed low-magnitude but identity-bearing activations, and the SCF in Eq. (6) zeros exactly those activations. The ablations in Tables 2-3 and the LayerCAM visualizations in Fig. 7 measure only end-to-end accuracy or coarse attention; they do not measure how many true-positive foreground locations are zeroed by the filter, nor whether the suppression rate differs between source and target domains. Please add per-layer statistics (e.g., the fraction of foreground versus background activations zeroed, or recall of pseudo-boxes before and after filtering) and, ideally, a control that applies the same nonlinearity with shuffled or fixed thresholds to rule out that PDAF acts merely as an unstructured regularizer.
  2. [§4.2, Table 2] Table 2 is not self-contained: rows use blank cells to indicate which component is removed, and the text refers to rows as #1-#4 while the table has no row numbers. As printed, the reader cannot reliably map each row to the corresponding ablated component, which undermines the ablation evidence for PDT, SCF, and CPR. Please replace the blank cells with explicit ✓/✗ markers and add row labels or row numbers, and correct the comma decimal in the 34.7 entry.
  3. [§4.1 and §4.3, Table 1] The headline improvement over the runner-up is 0.7 mAP on CUHK-SYSU and 3.2 mAP on PRW, but all results come from a single run with no error bars, and the epoch schedules are chosen per target dataset (7 pre-training + 13 joint epochs for PRW, 2 + 7 for CUHK-SYSU). This makes it difficult to distinguish genuine improvement from run-to-run variance or benchmark-specific tuning. Please report multiple seeds (at least three) with means and standard deviations, and either fix a shared schedule or provide a sensitivity analysis over epochs and the main hyperparameters (γ, m, and the contrastive temperature).
  4. [§3.3, Eq. (9)] The online cluster proxy update is written as ck ← −(1−γ)ck + γf. If taken literally, the negative sign makes the stored proxy alternate in sign with each update and is inconsistent with the momentum-update description in the text and with the positive combination in Eq. (10). Please confirm the intended formula and correct the sign in Eq. (9), since this equation is load-bearing for the CPR module.
minor comments (5)
  1. [Throughout] There are several naming inconsistencies: "DSAC" appears in the Method section, "DCSA" appears in the Experiments section, and the Introduction refers to "The SCT" where the Self-Calibrating Filter (SCF) is meant. Please unify the acronyms.
  2. [§3.2, Eq. (4)] The phrase "panning operation" for the soft threshold should be "shifting" or "shrinking operation" to describe the translation of the input by τ.
  3. [§3.2, Eqs. (5)-(6)] The order n of the High-Order Soft Threshold is never stated in the experimental setup; Fig. 3 only says "take n = 2 as an example." Please state the value of n used in all experiments and include an ablation over n, since the text motivates SCT as a remedy for gradient issues at larger n but never reports how n was chosen.
  4. [§4.2, Fig. 7 and Fig. 6] The qualitative evidence would be easier to judge if Fig. 6 had labeled axes with units, and if Fig. 7 showed the input image and attention map for the same scale in each row; the current composites are hard to compare.
  5. [§4.3, Table 1] The phrase "comparable to or even exceeding the performance of some fully supervised methods" is supported only against older supervised baselines such as OIM; the modern supervised methods (SeqNeXt, SEAS) remain clearly ahead. Please rephrase the claim to avoid implying competitiveness with current fully supervised person search.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the reported gains are empirical benchmark results with ablations, and the background self-citations are not load-bearing.

full rationale

The paper makes no first-principles derivation claim. Its central claims are architectural: PDAF filters backbone features using a threshold tau = alpha x AvgPool(F) (Eq. 2) and SCF (Eq. 6), and CPR replaces instance-level memory with cluster proxies (Eqs. 7-10). None of these equations assumes the reported mAP/top-1 numbers; they are defined independently and evaluated on external benchmarks (CUHK-SYSU, PRW) with ablations in Tables 2-3. The 'prediction' language is ordinary empirical comparison, not a fitted value renamed as a prediction. The measured performance is computed against held-out test annotations, not against the pseudo-labels used in training. The only self-citations are background references (Peng et al. 2020; Wang et al. 2022; Jiang et al. 2024; Wang et al. 2024a) used to situate ReID and person search; they do not supply a uniqueness theorem or carry the method's load. The skeptical concern about PDAF thresholds suppressing target-domain foreground features is a correctness/robustness issue, not circularity: even if the filter were harmful, the components are not defined in terms of the outcome. Therefore no circular step is exhibited and the score is 0.

Assumptions & free parameters 7 free parameters · 6 assumptions · 0 invented entities

Most load-bearing choices are inherited from the DAPS pipeline: pretrained ResNet-50, RPN proposals as pseudo boxes, DBSCAN pseudo identities, and DAM alignment. The paper adds a heuristic channel-threshold filter and a cluster-proxy memory, both relying on the assumption that noisy pseudo-labels can be filtered without removing identity information. No new physical entities are introduced. The listed hyperparameters are tuned on the target benchmarks.

free parameters (7)
  • alpha (adaptive scaling factor from MLP) = learned
    Produced by sigmoid of MLP(MaxPool(F)) in Eq. 2; controls the adaptive threshold.
  • lambda (SCF blend weight) = learned, initialized to 0
    Eq. 6 blends the soft threshold and the high-order soft threshold; no final value is reported.
  • n (HST order) = 2
    Eq. 5; the authors state larger n is theoretically better but choose 2 to avoid gradient problems.
  • gamma (cluster proxy momentum) = 0.2
    Eq. 9 controls the online cluster proxy update speed; set by hand.
  • m (EMA smoothing factor) = 0.2
    Eq. 10 controls the offline smoothing of cluster proxies; set by hand.
  • contrastive temperature tau = not reported
    Eq. 8 uses tau as the softmax temperature; its value is not specified in the paper.
  • per-target epoch schedule = PRW target: 7 pretrain + 13 joint; CUHK-SYSU target: 2 pretrain + 7 joint
    Implementation Details says epochs are adjusted based on variations in the target domain dataset.
assumptions (6)
  • domain assumption ImageNet-pretrained ResNet-50 features provide a usable starting point for both source and target domains.
    Used as the backbone in Implementation Details; this transfer assumption is not tested.
  • domain assumption RPN-generated boxes on the target domain can serve as pseudo bounding boxes without additional correction.
    Cluster Labeling Method uses training proposals directly as pseudo bounding boxes.
  • domain assumption DBSCAN clustering of instance features produces pseudo identities whose cluster centroids are reliable proxies.
    Eq. 7 initializes cluster proxies from mean features; if clusters are bad, the contrastive loss in Eq. 8 trains on wrong identities.
  • domain assumption The DAM implicit domain alignment from DAPS remains effective inside the new DSCA framework.
    Framework overview builds on DAPS and keeps DAM without new evidence for its validity.
  • ad hoc to paper Channel-wise mean and max pooling statistics separate foreground from background in backbone feature maps.
    Eq. 2 defines the adaptive threshold from pooling statistics; no independent validation is given that this threshold preserves identity features.
  • ad hoc to paper Single clustering per training epoch is sufficient to maintain reliable pseudo-labels.
    The introduction claims one clustering per epoch versus multiple clustering in prior work, but no cluster-quality comparison is provided.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Unsupervised Domain Adaptive Person Search via Dual Self-Calibration." pith.science (2026). https://pith.science/paper/SYKDAT5V

@misc{pith2026241216506,
  author       = {Pith},
  title        = {Pith review of: Unsupervised Domain Adaptive Person Search via Dual Self-Calibration},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/SYKDAT5V}},
  note         = {Machine review of arXiv:2412.16506}
}
read the original abstract

Unsupervised Domain Adaptive (UDA) person search focuses on employing the model trained on a labeled source domain dataset to a target domain dataset without any additional annotations. Most effective UDA person search methods typically utilize the ground truth of the source domain and pseudo-labels derived from clustering during the training process for domain adaptation. However, the performance of these approaches will be significantly restricted by the disrupting pseudo-labels resulting from inter-domain disparities. In this paper, we propose a Dual Self-Calibration (DSCA) framework for UDA person search that effectively eliminates the interference of noisy pseudo-labels by considering both the image-level and instance-level features perspectives. Specifically, we first present a simple yet effective Perception-Driven Adaptive Filter (PDAF) to adaptively predict a dynamic filter threshold based on input features. This threshold assists in eliminating noisy pseudo-boxes and other background interference, allowing our approach to focus on foreground targets and avoid indiscriminate domain adaptation. Besides, we further propose a Cluster Proxy Representation (CPR) module to enhance the update strategy of cluster representation, which mitigates the pollution of clusters from misidentified instances and effectively streamlines the training process for unlabeled target domains. With the above design, our method can achieve state-of-the-art (SOTA) performance on two benchmark datasets, with 80.2% mAP and 81.7% top-1 on the CUHK-SYSU dataset, with 39.9% mAP and 81.6% top-1 on the PRW dataset, which is comparable to or even exceeds the performance of some fully supervised methods. Our source code is available at https://github.com/whbdmu/DSCA.

Figures

Figures reproduced from arXiv: 2412.16506 by the authors.

Figure 1
Figure 1. Noisy pseudo-labels consist of low-quality pseudo [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. The design architecture of the DSCA framework. For each training period, DSCA alternates between two phases:(1) [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Comparison of our High-Order Soft Threshold [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Visualization of the detection scores and re [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]
Figure 5
Figure 5. Figure 5: Qualitative comparison of DSCA with DAPS on [PITH_FULL_IMAGE:figures/full_fig_p006_5.png]
Figure 6
Figure 6. Figure 6: Number of cluster representations stored in the [PITH_FULL_IMAGE:figures/full_fig_p007_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

43 extracted references · 36 canonical work pages

  1. [1]

    , " * write output.state after.block = add.period write newline

    ENTRY address archivePrefix author booktitle chapter edition editor eid eprint howpublished institution isbn journal key month note number organization pages publisher school series title type volume year label extra.label sort.label short.list INTEGERS output.state before.all mid.sentence after.sentence after.block FUNCTION init.state.consts #0 'before.a...

  2. [2]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION word.in bbl.in capitalize " " * FUNCT...

  3. [3]

    K.; Fiaz, M.; and Cholakkal, H

    Almansoori, M. K.; Fiaz, M.; and Cholakkal, H. 2024. DDAM-PS: Diligent Domain Adaptive Mixer for Person Search. In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, 6688--6697

  4. [4]

    M.; Cholakkal, H.; Xie, J.; Shah, M.; and Khan, F

    Cao, J.; Pang, Y.; Anwer, R. M.; Cholakkal, H.; Xie, J.; Shah, M.; and Khan, F. S. 2022. PSTR: End-to-End One-Step Person Search With Transformers. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 9458--9467

  5. [5]

    Chen, D.; Zhang, S.; Ouyang, W.; Yang, J.; and Schiele, B. 2020 a . Hierarchical online instance matching for person search. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 34, 10518--10525

  6. [6]

    Chen, D.; Zhang, S.; Ouyang, W.; Yang, J.; and Tai, Y. 2018 a . Person search via a mask-guided two-stream cnn model. In Proceedings of the european conference on computer vision (ECCV), 734--750

  7. [7]

    Chen, D.; Zhang, S.; Yang, J.; and Schiele, B. 2020 b . Norm-aware embedding for efficient person search. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 12615--12624

  8. [8]

    Chen, Y.; Li, W.; Sakaridis, C.; Dai, D.; and Van Gool, L. 2018 b . Domain adaptive faster r-cnn for object detection in the wild. In Proceedings of the IEEE conference on computer vision and pattern recognition, 3339--3348

Show all 43 references
  1. [9]

    Dai, Z.; Wang, G.; Yuan, W.; Zhu, S.; and Tan, P. 2022. Cluster contrast for unsupervised person re-identification. In Proceedings of the Asian conference on computer vision, 1142--1160

  2. [10]

    Deng, W.; Zheng, L.; Ye, Q.; Kang, G.; Yang, Y.; and Jiao, J. 2018. Image-image domain adaptation with preserved self-similarity and domain-dissimilarity for person re-identification. In Proceedings of the IEEE conference on computer vision and pattern recognition, 994--1003

  3. [11]

    Donoho, D. L. 1995. De-noising by soft-thresholding. IEEE transactions on information theory, 41(3): 613--627

  4. [12]

    L.; and Johnstone, I

    Donoho, D. L.; and Johnstone, I. M. 1994. Ideal spatial adaptation by wavelet shrinkage. biometrika, 81(3): 425--455

  5. [13]

    Ester, M.; Kriegel, H.-P.; Sander, J.; Xu, X.; et al. 1996. A density-based algorithm for discovering clusters in large spatial databases with noise. In kdd, volume 96, 226--231

  6. [14]

    Ganin, Y.; and Lempitsky, V. 2015. Unsupervised domain adaptation by backpropagation. In International conference on machine learning, 1180--1189. PMLR

  7. [15]

    Ge, Y.; Zhu, F.; Chen, D.; Zhao, R.; et al. 2020. Self-paced contrastive learning with hybrid memory for domain adaptive object re-id. Advances in neural information processing systems, 33: 11309--11321

  8. [16]

    Girshick, R.; Donahue, J.; Darrell, T.; and Malik, J. 2014. Rich feature hierarchies for accurate object detection and semantic segmentation. In Proceedings of the IEEE conference on computer vision and pattern recognition, 580--587

  9. [17]

    Han, C.; Su, K.; Yu, D.; Yuan, Z.; Gao, C.; Sang, N.; Yang, Y.; and Wang, C. 2021. Weakly supervised person search with region siamese networks. In Proceedings of the IEEE/CVF International Conference on Computer Vision, 12006--12015

  10. [18]

    Han, C.; Ye, J.; Zhong, Y.; Tan, X.; Zhang, C.; Gao, C.; and Sang, N. 2019. Re-id driven localization refinement for person search. In Proceedings of the IEEE/CVF International Conference on Computer Vision, 9814--9823

  11. [19]

    He, K.; Fan, H.; Wu, Y.; Xie, S.; and Girshick, R. 2020. Momentum contrast for unsupervised visual representation learning. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 9729--9738

  12. [20]

    He, K.; Zhang, X.; Ren, S.; and Sun, J. 2016. Deep residual learning for image recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition, 770--778

  13. [21]

    Huang, Z.; Ren, Y.; Pu, X.; Huang, S.; Xu, Z.; and He, L. 2023. Self-supervised graph attention networks for deep weighted multi-view clustering. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 37, 7936--7943

  14. [22]

    Jaffe, L.; and Zakhor, A. 2023. Gallery filter network for person search. In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, 1684--1693

  15. [23]

    Jiang, P.-T.; Zhang, C.-B.; Hou, Q.; Cheng, M.-M.; and Wei, Y. 2021. Layercam: Exploring hierarchical class activation maps for localization. IEEE Transactions on Image Processing, 30: 5875--5888

  16. [24]

    Jiang, Y.; Wang, H.; Peng, J.; Fu, X.; and Wang, Y. 2024. Scene-Adaptive Person Search via Bilateral Modulations. In Proceedings of the Thirty-Third International Conference on International Joint Conferences on Artificial Intelligence

  17. [25]

    Kang, G.; Jiang, L.; Yang, Y.; and Hauptmann, A. G. 2019. Contrastive adaptation network for unsupervised domain adaptation. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 4893--4902

  18. [26]

    Li, J.; Yan, Y.; Wang, G.; Yu, F.; Jia, Q.; and Ding, S. 2022. Domain adaptive person search. In European Conference on Computer Vision, 302--318. Springer

  19. [27]

    Li, Z.; and Miao, D. 2021. Sequential end-to-end network for efficient person search. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 35, 2011--2019

  20. [28]

    Liu, H.; Wang, Y.; Qian, B.; Wang, M.; and Rui, Y. 2024. Structure Matters: Tackling the Semantic Discrepancy in Diffusion Models for Image Inpainting. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 8038--8047

  21. [29]

    Peng, J.; Wang, Y.; Wang, H.; Zhang, Z.; Fu, X.; and Wang, M. 2020. Unsupervised vehicle re-identification with progressive adaptation. arXiv preprint arXiv:2006.11486

  22. [30]

    Ren, S.; He, K.; Girshick, R.; and Sun, J. 2015. Faster r-cnn: Towards real-time object detection with region proposal networks. Advances in neural information processing systems, 28

  23. [31]

    Wang, B.; Yang, Y.; Wu, J.; Qi, G.-j.; and Lei, Z. 2023. Self-similarity driven scale-invariant learning for weakly supervised person search. In Proceedings of the IEEE/CVF International Conference on Computer Vision, 1813--1822

  24. [32]

    Wang, C.; Ma, B.; Chang, H.; Shan, S.; and Chen, X. 2020. Tcts: A task-consistent two-stage framework for person search. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 11952--11961

  25. [33]

    Wang, H.; Yao, M.; Chen, Y.; Xu, Y.; Liu, H.; Jia, W.; Fu, X.; and Wang, Y. 2024 a . Manifold-based Incomplete Multi-view Clustering via Bi-Consistency Guidance. IEEE Transactions on Multimedia

  26. [34]

    Wang, J.; Pang, Y.; Cao, J.; Sun, H.; Shao, Z.; and Li, X. 2024 b . Deep intra-image contrastive learning for weakly supervised one-step person search. Pattern Recognition, 147: 110047

  27. [35]

    Wang, Y.; Peng, J.; Wang, H.; and Wang, M. 2022. Progressive learning with multi-scale attention network for cross-domain vehicle re-identification. Science China Information Sciences, 65(6): 160103

  28. [36]

    Wei, X.-S.; Song, Y.-Z.; Mac Aodha, O.; Wu, J.; Peng, Y.; Tang, J.; Yang, J.; and Belongie, S. 2021. Fine-grained image analysis with deep learning: A survey. IEEE transactions on pattern analysis and machine intelligence, 44(12): 8927--8948

  29. [37]

    Xiao, T.; Li, S.; Wang, B.; Lin, L.; and Wang, X. 2017. Joint detection and identification feature learning for person search. In Proceedings of the IEEE conference on computer vision and pattern recognition, 3415--3424

  30. [38]

    Yan, Y.; Li, J.; Liao, S.; Qin, J.; Ni, B.; Lu, K.; and Yang, X. 2022. Exploring visual context for weakly supervised person search. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 36, 3027--3035

  31. [39]

    Yan, Y.; Li, J.; Qin, J.; Bai, S.; Liao, S.; Liu, L.; Zhu, F.; and Shao, L. 2021. Anchor-free person search. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 7690--7699

  32. [40]

    Ye, M.; Lan, X.; Leng, Q.; and Shen, J. 2020. Cross-modality person re-identification via modality-aware collaborative ensemble learning. IEEE Transactions on Image Processing, 29: 9387--9399

  33. [41]

    Ye, M.; Wu, Z.; Chen, C.; and Du, B. 2023. Channel augmentation for visible-infrared re-identification. IEEE Transactions on Pattern Analysis and Machine Intelligence

  34. [42]

    Zhao, M.; Zhong, S.; Fu, X.; Tang, B.; and Pecht, M. 2019. Deep residual shrinkage networks for fault diagnosis. IEEE Transactions on Industrial Informatics, 16(7): 4681--4690

  35. [43]

    Zheng, L.; Zhang, H.; Sun, S.; Chandraker, M.; Yang, Y.; and Tian, Q. 2017. Person re-identification in the wild. In Proceedings of the IEEE conference on computer vision and pattern recognition, 1367--1376

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.