Pith. sign in

REVIEW 3 major objections 6 minor 56 references

Collaborative Enhancement Network for Low-quality Multi-spectral Vehicle Re-identification

T0 review · 3 major / 6 minor · reviewed 2026-08-16 · deepseek-v4-flash

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

Pith's one-line read CoEN fuses RGB, near-infrared, and thermal spectra into a proxy that selects the best primary spectrum, lifting multi-spectral vehicle re-identification accuracy on three benchmarks.

desk verdict A plausible incremental method with a genuinely new dynamic primary-selection idea, but the mechanism is under-validated and the headline gains lack error bars; worth peer review, not a desk reject. read the letter →

arxiv 2504.14877 v1 pith:JSE2YTMS submitted 2025-04-21 cs.CV

Aihua Zheng , Yongqi Sun , Zi Wang , Chenglong Li , Jin Tang This is my paper · ORCID
classification cs.CV
keywords multi-spectralvehiclere-identificationlow-qualityspectraproxyfeaturegenerationdynamicqualitysortcollaborativeenhancementcross-attentionVisionTransformerflareandlow-illuminationscenes
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 proposes Collaborative Enhancement Network (CoEN) for re-identifying vehicles from RGB, near-infrared, and thermal images when some spectra are degraded by flares, low light, or other conditions. The central claim is that instead of trusting a fixed primary spectrum, the network should build a fused proxy from all spectra, use that proxy to score which spectrum currently carries the most identity information, and then enhance every spectrum using both the best spectrum and the proxy. On the flare-heavy WMVeID863 benchmark, CoEN reports 71.4% mAP and 79.2% Rank-1, exceeding the prior FACENet by 1.6 and 2.2 points respectively, and on RGBNT100 and MSVR310 it reports larger gains over the previous best DeMo. The intended contribution is a scene-adaptive answer to the question of which spectrum should lead the enhancement, rather than a hand-picked prior.

What carries the argument

The central object is the proxy feature $F_P = \mathrm{ProjB}(\mathrm{ProjA}(\mathrm{Concat}(F_R^{emb}, F_N^{emb}, F_T^{emb})))$, a fused representation formed by concatenating the embedding tokens of all three spectra and reducing their dimension in two steps. The proxy acts as a shared currency: the Dynamic Quality Sort Module uses token-level cosine similarities between $F_P$ and each spectrum's embedding to assign quality scores and rank the spectra, while the Collaborative Enhancement Module uses $F_P$ as key and value in a proxy-based cross-attention branch. The same proxy is also used as a branch for final identity inference. This design replaces any fixed choice of primary spectrum with a per-sample, proxy-driven selection.

What would settle it

Record the primary spectrum chosen by DQSM for each sample on WMVeID863 and check whether the choice varies across flare, normal, and low-light scenes; if the selected primary is almost always TIR, or if the DQSM ranking does not change when the proxy is computed without the spectrum being scored, then the dynamic quality sort is not driving the reported gains.

Watch

Extended reading notes

Core claim

The paper claims that low-quality multi-spectral vehicle re-identification is best served by a collaboratively enhanced representation, not by enhancing all spectra from one presumed-good spectrum. CoEN builds a proxy feature by progressively fusing embedding tokens from RGB, NIR, and TIR, then ranks spectra by the average cosine similarity of their tokens to the proxy and selects the top-ranked spectrum as primary. The Collaborative Enhancement Module applies cross-attention twice: primary-based enhancement transfers detailed cues from the primary to the two lower-ranked spectra, and proxy-based enhancement supplements every spectrum, including the primary, with fused discriminative cues; residual connections preserve spectrum-specific detail. With a shared Vision Transformer backbone and identity plus triplet losses, the paper reports state-of-the-art results across WMVeID863, RGBNT100, and MSVR310.

Load-bearing premise

The load-bearing premise is that a spectrum's similarity to the merged proxy tells you how much identity information that spectrum still has, even though the proxy was built from the very same spectra; if the similarity merely reflects the way the merging was constructed, the dynamic primary selection loses its advantage.

Editorial extensions

If this is right

  • If the reported gains hold, multi-spectral vehicle re-identification no longer needs a hand-picked primary spectrum such as thermal, because CoEN selects the primary per sample from the proxy ranking.
  • The two-pronged enhancement strategy (primary-based plus proxy-based cross-attention) is designed to work when all spectra are low quality, a setting where fixed-primary methods degrade, and the paper's ablations support this on the flare-heavy WMVeID863 benchmark.
  • Because the proxy itself carries more identity information than any single spectrum, the paper shows that adding the proxy as an inference branch improves retrieval even after enhancement, from 69.7% to 71.4% mAP on WMVeID863.
  • CoEN with a CLIP backbone also improves MSVR310 mAP to 52.2% versus 49.2% for DeMo, suggesting the collaborative enhancement transfers to large pre-trained backbones without requiring text annotations.
  • Ablation results indicate that progressive projection in the Proxy Generator outperforms element-wise summation or single-step projection, and that dynamic DQSM selection outperforms fixing RGB, NIR, or TIR as the primary spectrum.

Reading between the lines

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

  • A natural testable extension is to compare the DQSM's cosine-similarity ranking against ground-truth quality labels or no-reference image-quality scores per spectrum; if the ranking agrees with those labels better than a fixed thermal prior does, the dynamic selection claim is strengthened.
  • Because the proxy is built from the same spectra it scores, the quality scores may reflect which spectrum the fusion projection preserves rather than true scene-adapted quality; an editor's suggested experiment is to train the proxy on only two spectra and score the held-out third spectrum, checking whether the ranking still tracks known degradations.
  • The collaborative-enhancement design could plausibly transfer to other multi-modal recognition tasks, such as visible-infrared person re-identification or autonomous driving perception, where the best modality varies with scene conditions and a proxy could serve the same role as a scene-adaptive selector.
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

3 major / 6 minor

Summary. The paper proposes a Collaborative Enhancement Network (CoEN) for multi-spectral vehicle re-identification under low-quality, scene-varying degradations. A shared Vision Transformer extracts per-spectrum embedding tokens; the Proxy Generator (PG) concatenates these tokens and applies two progressive projections to form a fused proxy feature. The Dynamic Quality Sort Module (DQSM) ranks RGB, NIR, and TIR spectra by the mean cosine similarity between each spectrum's embedding tokens and the proxy, and selects the highest-scoring spectrum as the primary spectrum. The Collaborative Enhancement Module (CEM) then enhances the remaining spectra via two cross-attention branches, one using the primary spectrum as key/value and one using the proxy, with residual aggregation. Training uses identity and triplet losses on the backbone, CEM, and proxy branches. The paper reports state-of-the-art or competitive results on WMVeID863, RGBNT100, and MSVR310, together with ablations, hyperparameter analysis, and qualitative visualizations.

Significance. If the reported results hold, CoEN would be a practically useful state-of-the-art method on three public multi-spectral vehicle ReID benchmarks, and its central idea of making primary-spectrum selection scene-adaptive is a reasonable and timely research direction. The paper's strengths are its external comparisons to previously published methods, its internally consistent ablations in which each added component improves mAP, and its clear architectural description. The significance is currently conditional, however, because the paper's claimed novelty rests on DQSM's ability to measure spectral quality, and that measurement is not validated independently of the proxy that DQSM itself builds from the spectra being scored. The headline numerical gains over fixed TIR selection are also reported as single runs without error bars, so the magnitude of the dynamic-selection benefit is not yet established.

major comments (3)
  1. [§3.3, Eqs. (2)-(7), Table 6] The DQSM quality score is not an independent measure of spectral quality. The proxy F_P in Eq. (3) is generated by a learned projection of the concatenation of the very embeddings F_R, F_N, and F_T that Eqs. (5)-(6) then score against, so the correlation score Q_R measures agreement with a feature that is itself a function of F_R. If the projections are approximately linear, F_P is a mixture of the three spectra, and a spectrum can rank first because its contribution to the mixture or its per-token norm is largest, not because it retains more identity cues. The paper provides no ground-truth quality labels, no per-scene selection statistics, and no report of how often each spectrum is chosen as primary; Table 6 shows only a 0.8-1.0 mAP gap over fixed TIR selection with no error bars. The central 'accurate dynamic primary selection' claim is therefore unsupported as stated. A useful control would be to compare DQSM against random or fixed orderings with identical seeds, and to report selection frequencies per scene category; the authors should also clarify whether DQSM is intended as an intrinsic quality estimator or as a consensus-with-proxy measure.
  2. [§4.2, §4.5, Tables 1, 3, 6] The paper reports no error bars or multiple-seed statistics for any of its quantitative claims, and the two hyperparameters gamma (Eq. (9)) and lambda (Eq. (16)) are tuned on WMVeID863 and then evaluated on the same dataset. Table 1's 1.6% mAP advantage over FACENet and Table 6's 0.8% dynamic-selection advantage are of the same order as typical run-to-run variance in ReID experiments, so the current evidence does not exclude the possibility that the main margins are partly noise. I request at least three independent runs with mean and standard deviation for the principal tables, or an explicit evaluation protocol with a held-out validation split for hyperparameter selection. Without this, the ablations in Tables 3, 6, and 7 cannot be reliably used to attribute the gains to DQSM and CEM.
  3. [§3.3, Eq. (7) and §3.4, Eqs. (8)-(11)] The paper does not describe how the discrete sorting operation in Eq. (7) interacts with end-to-end training, nor whether the selected primary spectrum is stable across training epochs. If the ranking changes frequently for a given sample, the enhancement target of the primary-based branch is noisy, and the benefits of dynamic selection could be diluted. At minimum, the authors should include a training-time analysis of selection stability, and in the experiments report how often each spectrum is chosen as primary on WMVeID863 overall and separately for normal, flare, and low-illumination scenes. This would also address the question of whether DQSM is truly necessary or whether a fixed prior such as FACENet's TIR selection is sufficient in practice; the current Table 6 comparison is too coarse to resolve that question.
minor comments (6)
  1. [§4.1, first paragraph] The dataset name is spelled 'MWVeID863' at the start of Section 4.1; it should be 'WMVeID863' to match the rest of the paper.
  2. [Eq. (12)] The subscript '3nd' in Eq. (12) is a typo and should be '3rd'.
  3. [Figure 10 and §4.6] The text and figure caption refer to 'Demo' when the cited method is 'DeMo'; please unify the name.
  4. [§3.4, Eq. (9)] The dropout notation Dropout_γ is used before it is defined in the text; define gamma in a sentence preceding Eq. (9) or move the definition earlier.
  5. [Data Availability Statement] The Data Availability Statement says the data are included in the paper, but the experiments use third-party benchmark datasets; the statement should identify the public sources and access information instead.
  6. [Tables 1 and 2] The tables would be clearer if the multiple-seed/error-bar columns were added, and if the CLIP variant of CoEN on RGBNT100 were explicitly marked as not exceeding IDEA (87.2 mAP) in the caption, since the current text only describes it as competitive.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: benchmark gains are externally grounded, and the DQSM quality score, though self-referential as a heuristic, is explicitly defined as correlation to the proxy rather than derived from an independent quality label.

full rationale

The paper's central claims are empirical comparisons against external baselines on three public benchmarks (WMVeID863, RGBNT100, MSVR310), so the headline performance results are not derived from the method's own assumptions. The DQSM quality score is indeed self-referential: Eq. (5) defines Q_R as the mean cosine similarity between F_P and F_R, while Eqs. (2)-(3) construct F_P by projecting a concatenation of F_R, F_N, and F_T. Thus a spectrum can score highly partly because the proxy preserves its own content, and the paper does not validate the ranking against ground-truth quality labels. However, this is not a circular derivation in the formal sense: the paper explicitly defines the 'quality score' as this correlation, rather than claiming to derive an independent quality measure from first principles. The statement that a larger score means richer discriminative information is an unvalidated heuristic claim, which is a correctness/robustness concern, not a circularity. The paper also does not fit a parameter to a subset of data and then rename the fit as a prediction; the ablations in Tables 3-7 are end-to-end comparisons. Self-citations to FACENet and WMVeID863 are used as baselines and motivation, not as load-bearing justification for the proposed mechanism. No uniqueness theorem or ansatz is imported from the authors' prior work to forbid alternatives. Overall, the derivation chain is self-contained with respect to its empirical claims, and the possible weakness of the DQSM ranking does not make the paper's results equivalent to its inputs by construction.

Assumptions & free parameters 3 free parameters · 5 assumptions · 1 invented entities

The central claim is empirical and benchmark-grounded, so the ledger burden is moderate for a deep learning architecture paper. The only quantified free choices are two tuned scalars and a standard training schedule. The load-bearing conceptual entity is the proxy, whose twin roles as quality anchor and enhancement source rest on an untested cosine-similarity premise. The datasets and competing numbers come from prior work, including the same lab's FACENet, CCNet, and PFNet line, which the paper benchmarks against.

free parameters (3)
  • gamma (dropout rate in CEM cross-attention) = 0.5
    Set in Section 4.2 and tuned from 0.1 to 1.0 on WMVeID863 (Section 4.5, Fig. 5), the same benchmark used for the headline Table 1 numbers.
  • lambda (loss weight between ViT and CEM branches) = 0.5
    Set in Section 4.2 and tuned on WMVeID863 (Section 4.5, Fig. 6); directly affects the reported mAP.
  • Training schedule (SGD learning rate, epochs, batch composition) = 3e-3 on WMVeID863, 1e-3 on RGBNT100 and MSVR310; 120 epochs; batches of 4 identities x 8 samples
    Chosen by hand following prior ReID practice (Section 4.2); standard but not justified per task.
assumptions (5)
  • domain assumption ImageNet-pretrained DeiT-B/16 features transfer to RGB/NIR/TIR vehicle imagery.
    Section 4.2 initializes the shared ViT with ImageNet weights, assuming the pretrained feature space is useful for near-infrared and thermal input images, which are far outside ImageNet's distribution.
  • domain assumption A single shared ViT can represent all three spectra without modality-specific parameters.
    Section 3.1 shares one ViT across RGB, NIR, and TIR; the paper provides no experiment isolating whether weight sharing helps or hurts individual spectra.
  • domain assumption The fused proxy retains identity cues better than any single spectrum and is a reliable quality anchor.
    Section 3.2 asserts the proxy contains rich vehicle discriminative information; Table 5 shows proxy-only inference exceeds single-spectrum inference, but this is downstream evidence, not a guarantee for the quality-ranking role.
  • ad hoc to paper Mean token-wise cosine similarity between a spectrum and the proxy measures the spectrum's quality and remaining identity information.
    Section 3.3, Eqs. 4-7, defines quality scores exactly this way; the proxy is built from the scored spectra (Eqs. 2-3), and no ground-truth quality labels are used, so this premise is assumed rather than tested.
  • standard math Cross-entropy plus hard triplet loss with a fixed margin is an adequate objective for this ReID task.
    Section 3.5 adopts standard ReID losses from Szegedy et al. and Hermans et al. without task-specific justification; standard practice in the field.
invented entities (1)
  • Proxy feature F_P
    purpose: Unified multi-spectral representation that serves as the quality benchmark in DQSM, the key and value source in proxy-based enhancement, and an inference feature.
    F_P is a learned internal latent (Eqs. 2-3) with no observable or falsifiable handle outside the trained network; its utility is only measurable through downstream ReID accuracy, so independent_evidence is false.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Collaborative Enhancement Network for Low-quality Multi-spectral Vehicle Re-identification." pith.science (2026). https://pith.science/paper/JSE2YTMS

@misc{pith2026250414877,
  author       = {Pith},
  title        = {Pith review of: Collaborative Enhancement Network for Low-quality Multi-spectral Vehicle Re-identification},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/JSE2YTMS}},
  note         = {Machine review of arXiv:2504.14877}
}
read the original abstract

The performance of multi-spectral vehicle Re-identification (ReID) is significantly degraded when some important discriminative cues in visible, near infrared and thermal infrared spectra are lost. Existing methods generate or enhance missing details in low-quality spectra data using the high-quality one, generally called the primary spectrum, but how to justify the primary spectrum is a challenging problem. In addition, when the quality of the primary spectrum is low, the enhancement effect would be greatly degraded, thus limiting the performance of multi-spectral vehicle ReID. To address these problems, we propose the Collaborative Enhancement Network (CoEN), which generates a high-quality proxy from all spectra data and leverages it to supervise the selection of primary spectrum and enhance all spectra features in a collaborative manner, for robust multi-spectral vehicle ReID. First, to integrate the rich cues from all spectra data, we design the Proxy Generator (PG) to progressively aggregate multi-spectral features. Second, we design the Dynamic Quality Sort Module (DQSM), which sorts all spectra data by measuring their correlations with the proxy, to accurately select the primary spectra with the highest correlation. Finally, we design the Collaborative Enhancement Module (CEM) to effectively compensate for missing contents of all spectra by collaborating the primary spectra and the proxy, thereby mitigating the impact of low-quality primary spectra. Extensive experiments on three benchmark datasets are conducted to validate the efficacy of the proposed approach against other multi-spectral vehicle ReID methods. The codes will be released at https://github.com/yongqisun/CoEN.

Figures

Figures reproduced from arXiv: 2504.14877 by the authors.

Figure 1
Figure 1. Performance of different spectra in complex scenes. RGB and NIR spectra perform better in normal illumination scenes, TIR spectra contain more identity in￾formation in flare scenes, while NIR and TIR spectra show superior performance in low illumination scenes However, different spectra inevitably lead to the loss of important discriminative cues Zheng et al. (2025); Wang et al. (2025b) for vehicles due to their spe… view at source ↗
Figure 2
Figure 2. (a) Mutual-based enhancement method Wang et al. (2024, 2025b,a). (b) Specific spectral-based primary enhancement method Zheng et al. (2025). (c) Our method. These quality differences exacerbate domain discrep￾ancies between spectra, posing significant challenges for multi-spectral vehicle ReID. In real-world scenarios, low￾quality spectra often contain more noise, leading to more ambiguous and unstable vehicle featu… view at source ↗
Figure 3
Figure 3. The framework of our proposed Collaborative Enhancement Network (CoEN). First, features (𝐹𝑅 , 𝐹𝑁 , 𝐹𝑇 ) from different spectra are extracted using a ViT Dosovitskiy et al. (2020) backbone network with shared parameters. These features are fed into the Proxy Generator (PG) to generate fused proxy feature 𝐹𝑃 . Next, the Dynamic Quality Sort Module (DQSM) sorts and selects high-quality spectra feature 𝐹1st. Finally, th… view at source ↗
Figures from the paper (6 more)
Figure 4
Figure 4. Figure 4: Illustration of Collaborative Enhancement Module. in each spectrum by extracting rich discriminative cues from proxy and detailed information from high-quality primary spectra. Specifically, CEM comprises two com￾ponents: primary-based enhancement and proxy-based en￾ha…
Figure 5
Figure 5. Figure 5: Effect of 𝛾 on mAP (%) and Rank-1 (%) on WMVeID863 [PITH_FULL_IMAGE:figures/full_fig_p011_5.png]
Figure 6
Figure 6. Figure 6: Effect of 𝜆 on mAP (%) and Rank-1 (%) on WMVeID863 Frequency Baseline Distance CoEN Distance [PITH_FULL_IMAGE:figures/full_fig_p011_6.png]
Figure 7
Figure 7. Figure 7: The distributions of the two types of distances between the multi-spectral vehicle features on WMVeID863. excessively high or low values can cause the model to focus improperly on features, disrupting the equilibrium of the optimization process. The experimental result…
Figure 8
Figure 8. Figure 8: Comparison of feature distributions with T￾SNE Van der Maaten and Hinton (2008) on WMVeID863. Different colors represent different identities. Proxy-based Enhancemnet CoEN (Ours) Baseline Primary-based Enhancemnet [PITH_FULL_IMAGE:figures/full_fig_p012_8.png]
Figure 9
Figure 9. Figure 9: Class activation map visualization on WMVeID863 for different enhancement strategies. Feature Distributions. We employ the T-SNE Van der Maaten and Hinton (2008) dimensionality reduction tech￾nique to visualize and analyze the feature distributions of 15 vehicles. As s…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

56 extracted references · 47 canonical work pages

  1. [1]

    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 ":" * " " * FUNCTION f...

  2. [2]

    , author Hospedales, T.M

    author Chang, X. , author Hospedales, T.M. , author Xiang, T. , year 2018 . title Multi-level factorisation net for person re-identification , in: booktitle Proc. IEEE/CVF Conference on Computer Vision and Pattern Recognition , pp. pages 2109--2118

    work page 2018

  3. [3]

    , author Zhang, T

    author Chen, G. , author Zhang, T. , author Lu, J. , author Zhou, J. , year 2019 . title Deep meta metric learning , in: booktitle Proc. IEEE/CVF International Conference on Computer Vision , pp. pages 9547--9556

    work page 2019

  4. [4]

    , author Liu, C.T

    author Chen, T.S. , author Liu, C.T. , author Wu, C.W. , author Chien, S.Y. , year 2020 . title Orientation-aware vehicle re-identification with semantics-guided part attention network , in: booktitle Proc. European Conference on Computer Vision , pp. pages 330--346

    work page 2020

  5. [5]

    , author Chen, I

    author Chen, W. , author Chen, I. , author Yeh, C. , author Yang, H. , author Ding, J. , author Kuo, S. , year 2022 . title Sjdl-vehicle: Semi-supervised joint defogging learning for foggy vehicle re-identification , in: booktitle Proc. AAAI Conference on Artificial Intelligence , pp. pages 347--355

    work page 2022

  6. [6]

    , author Dong, W

    author Deng, J. , author Dong, W. , author Socher, R. , author Li, L.J. , author Li, K. , author Fei-Fei, L. , year 2009 . title Imagenet: A large-scale hierarchical image database , in: booktitle Proc. IEEE/CVF Conference on Computer Vision and Pattern Recognition , pp. pages 248--255

    work page 2009

  7. [7]

    , author Beyer, L

    author Dosovitskiy, A. , author Beyer, L. , author Kolesnikov, A. , author Weissenborn, D. , author Zhai, X. , author Unterthiner, T. , author Dehghani, M. , author Minderer, M. , author Heigold, G. , author Gelly, S. , et al., year 2020 . title An image is worth 16x16 words: Transformers for image recognition at scale . journal arXiv preprint arXiv:2010.11929

    arXiv 2020

  8. [8]

    , author Zhao, C

    author Guo, H. , author Zhao, C. , author Liu, Z. , author Wang, J. , author Lu, H. , year 2018 . title Learning coarse-to-fine structured feature embedding for vehicle re-identification , in: booktitle Proc. AAAI Conference on Artificial Intelligence

    work page 2018

Show all 56 references
  1. [9]

    , author Zhang, X

    author Guo, J. , author Zhang, X. , author Liu, Z. , author Wang, Y. , year 2022 . title Generative and attentive fusion for multi-spectral vehicle re-identification , in: booktitle International Conference on Intelligent Computing and Signal Processing , pp. pages 1565--1572 ...

    2022

  2. [10]

    , author Zhang, X

    author He, K. , author Zhang, X. , author Ren, S. , author Sun, J. , year 2016 . title Deep residual learning for image recognition , in: booktitle Proc. IEEE/CVF Conference on Computer Vision and Pattern Recognition , pp. pages 770--778

    2016

  3. [11]

    , author Lu, Z

    author He, Q. , author Lu, Z. , author Wang, Z. , author Hu, H. , year 2023 . title Graph-based progressive fusion network for multi-modality vehicle re-identification . journal IEEE Transactions on Intelligent Transportation Systems volume 24 , pages 12431--12447 . :10.1109/T...

    2023

  4. [12]

    , author Luo, H

    author He, S. , author Luo, H. , author Wang, P. , author Wang, F. , author Li, H. , author Jiang, W. , year 2021 . title Transreid: Transformer-based object re-identification , in: booktitle Proc. IEEE/CVF International Conference on Computer Vision , pp. pages 15013--15022

    2021

  5. [13]

    , author Beyer, L

    author Hermans, A. , author Beyer, L. , author Leibe, B. , year 2017 . title In defense of the triplet loss for person re-identification . journal arXiv preprint arXiv:1703.07737

    2017 arXiv

  6. [14]

    , author Liu, Z

    author Huang, G. , author Liu, Z. , author Van Der Maaten, L. , author Weinberger, K.Q. , year 2017 . title Densely connected convolutional networks , in: booktitle Proc. IEEE/CVF Conference on Computer Vision and Pattern Recognition , pp. pages 4700--4708

    2017

  7. [15]

    , author Xing, B

    author Huang, N. , author Xing, B. , author Zhang, Q. , author Han, J. , author Huang, J. , year 2024 . title Co-segmentation assisted cross-modality person re-identification . journal Information Fusion volume 104 , pages 102194

    2024

  8. [16]

    , author del Rinc\'on, J.M

    author Kamenou, E. , author del Rinc\'on, J.M. , author Miller, P. , author Devlin-Hill, P. , year 2023 . title A meta-learning approach for domain generalisation across visual modalities in vehicle re-identification , in: booktitle Proc. IEEE/CVF Conference on Computer Vision...

    2023

  9. [17]

    , author Chen, J

    author Li, H. , author Chen, J. , author Zheng, A. , author Wu, Y. , author Luo, Y. , year 2024 . title Day-night cross-domain vehicle re-identification , in: booktitle Proc. IEEE/CVF Conference on Computer Vision and Pattern Recognition , pp. pages 12626--12635

    2024

  10. [18]

    , author Li, C

    author Li, H. , author Li, C. , author Zheng, A. , author Tang, J. , author Luo, B. , year 2022 a. title Attribute and state guided structural embedding network for vehicle re-identification . journal IEEE Transactions on Image Processing volume 31 , pages 5949--5962

    2022

  11. [19]

    , author Li, C

    author Li, H. , author Li, C. , author Zhu, X. , author Zheng, A. , author Luo, B. , year 2020 . title Multi-spectral vehicle re-identification: A challenge , in: booktitle Proc. AAAI Conference on Artificial Intelligence , pp. pages 11345--11353

    2020

  12. [20]

    , author Ding, Z

    author Li, K. , author Ding, Z. , author Li, K. , author Zhang, Y. , author Fu, Y. , year 2022 b. title Vehicle and person re-identification with support neighbor loss . journal IEEE Transactions on Neural Networks and Learning Systems volume 33 , pages 826--838 . :10.1109/TNN...

    2022

  13. [21]

    , author Zhu, X

    author Li, W. , author Zhu, X. , author Gong, S. , year 2018 . title Harmonious attention network for person re-identification , in: booktitle Proc. IEEE/CVF Conference on Computer Vision and Pattern Recognition , pp. pages 2285--2294

    2018

  14. [22]

    , author Tian, Y

    author Liu, H. , author Tian, Y. , author Yang, Y. , author Pang, L. , author Huang, T. , year 2016 a. title Deep relative distance learning: Tell the difference between similar vehicles , in: booktitle Proc. IEEE/CVF Conference on Computer Vision and Pattern Recognition , pp....

    2016

  15. [23]

    , author Liu, W

    author Liu, X. , author Liu, W. , author Mei, T. , author Ma, H. , year 2016 b. title A deep learning-based approach to progressive vehicle re-identification for urban surveillance , in: booktitle Proc. European Conference on Computer Vision , pp. pages 869--884

    2016

  16. [24]

    , author Bai, Y

    author Lou, Y. , author Bai, Y. , author Liu, J. , author Wang, S. , author Duan, L. , year 2019 . title Veri-wild: A large dataset and a new method for vehicle re-identification in the wild , in: booktitle Proc. IEEE/CVF Conference on Computer Vision and Pattern Recognition ....

    2019

  17. [25]

    , author Gu, Y

    author Luo, H. , author Gu, Y. , author Liao, X. , author Lai, S. , author Jiang, W. , year 2019 . title Bag of tricks and a strong baseline for deep person re-identification , in: booktitle Proc. IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops

    2019

  18. [26]

    , author Hinton, G

    author Van der Maaten, L. , author Hinton, G. , year 2008 . title Visualizing data using t-sne . journal Journal of Machine Learning Research volume 9

    2008

  19. [27]

    , author Li, L

    author Meng, D. , author Li, L. , author Liu, X. , author Li, Y. , author Yang, S. , author Zha, Z.J. , author Gao, X. , author Wang, S. , author Huang, Q. , year 2020 . title Parsing-based view-aware embedding network for vehicle re-identification , in: booktitle Proc. IEEE/C...

    2020

  20. [28]

    , year 2023

    author Nianchang Huang, Jianan Liu, Y.M.Q.Z.J.H. , year 2023 . title Deep learning for visible-infrared cross-modality person re-identification: A comprehensive review . journal Information Fusion volume 91 , pages 396--411

    2023

  21. [29]

    , author Zhao, L

    author Pang, Z. , author Zhao, L. , author Liu, Y. , author Sharma, G. , author Wang, C. , year 2024 . title Inter-modality similarity learning for unsupervised multi-modality person re-identification . journal IEEE Transactions on Circuits and Systems for Video Technology

    2024

  22. [30]

    , author Cogswell, M

    author Selvaraju, R.R. , author Cogswell, M. , author Das, A. , author Vedantam, R. , author Parikh, D. , author Batra, D. , year 2020 . title Grad-cam: Visual explanations from deep networks via gradient-based localization , in: booktitle International Journal of Computer Vis...

    2020

  23. [31]

    , author Du, X

    author Shen, F. , author Du, X. , author Zhang, L. , author Shu, X. , author Tang, J. , year 2023 a. title Triplet contrastive representation learning for unsupervised vehicle re-identification . journal arXiv preprint arXiv:2301.09498

    2023 arXiv

  24. [32]

    , author Xie, Y

    author Shen, F. , author Xie, Y. , author Zhu, J. , author Zhu, X. , author Zeng, H. , year 2023 b. title Git: Graph interactive transformer for vehicle re-identification . journal IEEE Transactions on Image Processing volume 32 , pages 1039--1051

    2023

  25. [33]

    , author Cheng, C

    author Sun, Y. , author Cheng, C. , author Zhang, Y. , author Zhang, C. , author Zheng, L. , author Wang, Z. , author Wei, Y. , year 2020 . title Circle loss: A unified perspective of pair similarity optimization , in: booktitle Proc. IEEE/CVF Conference on Computer Vision and...

    2020

  26. [34]

    , author Zheng, L

    author Sun, Y. , author Zheng, L. , author Yang, Y. , author Tian, Q. , author Wang, S. , year 2018 . title Beyond part models: Person retrieval with refined part pooling (and a strong convolutional baseline) , in: booktitle Proc. European Conference on Computer Vision , pp. p...

    2018

  27. [35]

    , author Vanhoucke, V

    author Szegedy, C. , author Vanhoucke, V. , author Ioffe, S. , author Shlens, J. , author Wojna, Z. , year 2016 . title Rethinking the inception architecture for computer vision , in: booktitle Proc. IEEE/CVF Conference on Computer Vision and Pattern Recognition , pp. pages 2818--2826

    2016

  28. [36]

    , author Naphade, M

    author Tang, Z. , author Naphade, M. , author Liu, M.Y. , author Yang, X. , author Birchfield, S. , author Wang, S. , author Kumar, R. , author Anastasiu, D. , author Hwang, J.N. , year 2019 . title Cityflow: A city-scale benchmark for multi-target multi-camera vehicle trackin...

    2019

  29. [37]

    , author Cord, M

    author Touvron, H. , author Cord, M. , author Douze, M. , author Massa, F. , author Sablayrolles, A. , author Jégou, H. , year 2021 . title Training data-efficient image transformers distillation through attention . journal arXiv preprint arXiv:2012.12877

    2021 arXiv

  30. [38]

    , author Yuan, Y

    author Wang, G. , author Yuan, Y. , author Chen, X. , author Li, J. , author Zhou, X. , year 2018 . title Learning discriminative features with multiple granularities for person re-identification , in: booktitle Proc. ACM International Conference on Multimedia , pp. pages 274--282

    2018

  31. [39]

    , author Liu, H

    author Wang, T. , author Liu, H. , author Song, P. , author Guo, T. , author Shi, W. , year 2022 a. title Pose-guided feature disentangling for occluded person re-identification based on transformer , in: booktitle Proc. AAAI Conference on Artificial Intelligence , pp. pages 2...

    2022

  32. [40]

    , author Liu, X

    author Wang, Y. , author Liu, X. , author Yan, T. , author Liu, Y. , author Zheng, A. , author Zhang, P. , author Lu, H. , year 2025 a. title Mambapro: Multi-modal object re-identification with mamba aggregation and synergistic prompt , in: booktitle Proc. AAAI Conference on A...

    2025

  33. [41]

    , author Liu, X

    author Wang, Y. , author Liu, X. , author Zhang, P. , author Lu, H. , author Tu, Z. , author Lu, H. , year 2024 . title Top-reid: Multi-spectral object re-identification with token permutation , in: booktitle Proc. AAAI Conference on Artificial Intelligence

    2024

  34. [42]

    , author Liu, Y

    author Wang, Y. , author Liu, Y. , author Zheng, A. , author Zhang, P. , year 2025 b. title Demo: Decoupled feature-based mixture of experts for multi-modal object re-identification , in: booktitle Proc. AAAI Conference on Artificial Intelligence

    2025

  35. [43]

    , author Lv, Y

    author Wang, Y. , author Lv, Y. , author Zhang, P. , author Lu, H. , year 2025 c. title Idea: Inverted text with cooperative deformable aggregation for multi-modal object re-identification , in: booktitle Proc. IEEE/CVF Conference on Computer Vision and Pattern Recognition

    2025

  36. [44]

    , author Li, C

    author Wang, Z. , author Li, C. , author Zheng, A. , author He, R. , author Tang, J. , year 2022 b. title Interact, embed, and enlarge (ieee): Boosting modality-specific representations for multi-modal person re-identification , in: booktitle Proc. AAAI Conference on Artificia...

    2022

  37. [45]

    , author Shen, J

    author Ye, M. , author Shen, J. , author Lin, G. , author Xiang, T. , author Shao, L. , author Hoi, S.C. , year 2021 . title Deep learning for person re-identification: A survey and outlook . journal IEEE Transactions on Pattern Analysis and Machine Intelligence volume 44 , pa...

    2021

  38. [46]

    , author Wang, Y

    author Zhang, P. , author Wang, Y. , author Liu, Y. , author Tu, Z. , author Lu, H. , year 2024 . title Magic tokens: Select diverse tokens for multi-modal object re-identification , in: booktitle Proc. IEEE/CVF Conference on Computer Vision and Pattern Recognition

    2024

  39. [47]

    , author Zhou, X

    author Zhang, X. , author Zhou, X. , author Lin, M. , author Sun, J. , year 2018 . title Shufflenet: An extremely efficient convolutional neural network for mobile devices , in: booktitle Proc. IEEE/CVF Conference on Computer Vision and Pattern Recognition , pp. pages 6848--6856

    2018

  40. [48]

    , author Qi, F

    author Zhao, J. , author Qi, F. , author Ren, G. , author Xu, L. , year 2021 a. title Phd learning: Learning with pompeiu-hausdorff distances for video-based vehicle re-identification , in: booktitle Proc. IEEE/CVF Conference on Computer Vision and Pattern Recognition , pp. pa...

    2021

  41. [49]

    , author Zhao, Y

    author Zhao, J. , author Zhao, Y. , author Li, J. , author Yan, K. , author Tian, Y. , year 2021 b. title Heterogeneous relational complement for vehicle re-identification , in: booktitle Proc. IEEE/CVF International Conference on Computer Vision , pp. pages 205--214

    2021

  42. [50]

    , author Liu, J

    author Zheng, A. , author Liu, J. , author Wang, Z. , author Huang, L. , author Li, C. , author Yin, B. , year 2023 a. title Visible-infrared person re-identification via specific and shared representations learning . journal Visual Intelligence volume 1 , pages 29

    2023

  43. [51]

    , author Ma, Z

    author Zheng, A. , author Ma, Z. , author Sun, Y. , author Wang, Z. , author Li, C. , author Tang, J. , year 2025 . title Flare-aware cross-modal enhancement network for multi-spectral vehicle re-identification . journal Information Fusion volume 116 , pages 102800

    2025

  44. [52]

    , author Wang, Z

    author Zheng, A. , author Wang, Z. , author Chen, Z. , author Li, C. , author Tang, J. , year 2021 . title Robust multi-modality person re-identification , in: booktitle Proc. AAAI Conference on Artificial Intelligence , pp. pages 3529--3537

    2021

  45. [53]

    , author Zhu, X

    author Zheng, A. , author Zhu, X. , author Ma, Z. , author Li, C. , author Tang, J. , author Ma, J. , year 2023 b. title Cross-directional consistency network with adaptive layer normalization for multi-spectral vehicle re-identification and a high-quality benchmark . journal ...

    2023

  46. [54]

    , author Huang, X

    author Zheng, X. , author Huang, X. , author Ji, C. , author Yang, X. , author Sha, P. , author Cheng, L. , year 2024 . title Multi-modal person re-identification based on transformer relational regularization . journal Information Fusion volume 103 , pages 102128

    2024

  47. [55]

    , author Khosla, A

    author Zhou, B. , author Khosla, A. , author Lapedriza, A. , author Oliva, A. , author Torralba, A. , year 2016 . title Learning deep features for discriminative localization , in: booktitle Proc. IEEE/CVF International Conference on Computer Vision , pp. pages 2921--2929

    2016

  48. [56]

    , author Yang, Y

    author Zhou, K. , author Yang, Y. , author Cavallaro, A. , author Xiang, T. , year 2019 . title Omni-scale feature learning for person re-identification , in: booktitle Proc. IEEE/CVF International Conference on Computer Vision , pp. pages 3702--3712

    2019

Pith tools

Reviewed August 16, 2026 · model on record in the stance chip above.