AG-VPReID.VIR: Bridging Aerial and Ground Platforms for Video-based Visible-Infrared Person Re-ID

Akila Pemasiri; Akmal Jahan; Clinton Fookes; Huy Nguyen; Kien Nguyen; Sridha Sridharan

REVIEW 4 major objections 5 minor 55 references

AG-VPReID.VIR: Bridging Aerial and Ground Platforms for Video-based Visible-Infrared Person Re-ID

T0 review · 4 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read The paper introduces AG-VPReID.VIR, the first video-based person re-identification dataset combining aerial and ground platforms with visible and infrared modalities, along with a three-stream method that outperforms prior approaches by…

desk verdict The dataset is a genuinely useful first, but the method's headline gains are inflated by test-set hyperparameter selection, so treat the benchmark numbers with caution. read the letter →

arxiv 2507.17995 v1 pith:6PKJCYDE submitted 2025-07-24 cs.CV

Huy Nguyen , Kien Nguyen , Akila Pemasiri , Akmal Jahan , Clinton Fookes , Sridha Sridharan This is my paper · ORCID

classification cs.CV

keywords personre-identificationvisible-infraredaerial-groundsurveillanceUAVimagerythermalimagingvideo-basedcross-modalitybenchmarkdroneandCCTV

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

Person re-identification systems need to match the same person across cameras, but existing RGB-infrared benchmarks only cover ground-level views, and existing drone-based benchmarks only cover visible light. This paper introduces AG-VPReID.VIR, a video dataset of 1,837 identities in 4,861 tracklets and 124,855 frames, where the same people appear in RGB and infrared images from UAVs, fixed CCTV cameras, and wearable cameras. The authors argue that this combination—cross-platform and cross-modality at once—creates new and harder failure modes, and they report that established methods drop by more than half their accuracy on it. They also propose TCC-VPReID, a three-stream architecture that combines style-robust features, memory-based cross-view adaptation, and an intermediary-guided temporal stream, and report Rank-1/mAP of 36.18/41.56 for infrared-to-visible and 46.33/59.23 for visible-to-infrared ground-to-ground retrieval, more than 20 points above prior methods.

What carries the argument

The load-bearing object is the AG-VPReID.VIR dataset: 4,861 manually-reviewed tracklets of 1,837 identities, each identity labeled consistently across RGB and IR images from UAV, CCTV, and wearable cameras, with synchronized dual-sensor drone captures producing the first aerial infrared imagery for person re-identification. The method that carries the experimental claim is TCC-VPReID, a three-stream architecture: Stream 1 applies random channel-style perturbations plus intra-modal style attacks to force style-robust features; Stream 2 maintains separate aerial and ground memory prototypes per identity and aligns video features to them with a contrastive loss; Stream 3 converts RGB and IR frames into edge-based anaglyph images that serve as a modality-neutral intermediary, with cross-reconstruction pulling the two modalities together. The three streams are fused and trained with a weighted sum of identity, triplet, style-attack, cross-reconstruction, and video-to-memory losses.

What would settle it

Re-annotate a random sample of tracklets—say 200 identities spanning all five camera views—by independent annotators without access to the original labels, and measure per-frame and per-tracklet identity agreement; if agreement falls below a high threshold (e.g., 95%) where the original labels disagree, the benchmark's numbers cannot be trusted. A cheaper partial check is to train a simple baseline on a test split where labels are manually verified and see whether the reported 20-point gains persist.

Watch

Extended reading notes

Core claim

The paper's central claim is that person re-identification can and should be studied jointly across platforms (aerial and ground) and modalities (visible and infrared), and that AG-VPReID.VIR is the first dataset to make that possible. The dataset's unique value is aerial infrared imagery, which is harder than ground infrared because of small, low-resolution, top-down views of people. To show the setting is learnable, the paper builds TCC-VPReID, whose three streams each target one gap: the first attacks style variation with augmented channel perturbations, the second stores separate aerial and ground prototypes per identity in a memory bank with a video-to-memory contrastive loss, and the third uses edge-detected anaglyph images as a modality-neutral intermediary with cross-reconstruction. On the new benchmark the method lifts Rank-1 by more than 20 points over the best prior work, and on two existing ground-only video benchmarks it also improves on the previous state of the art.

Load-bearing premise

The dataset labels—the identity of every person in every RGB and IR tracklet, across aerial, CCTV, and wearable views—must be correct and consistent; if any tracklet mixes identities or any visible and infrared label is misaligned, every accuracy number on AG-VPReID.VIR is compromised.

Editorial extensions

If this is right

If the dataset is sound, it becomes the standard benchmark for aerial-ground RGB-IR video re-identification, letting researchers compare methods under a realistic 24-hour surveillance mix of drones, CCTV, and body cameras.
The reported sharp performance drop of existing methods when moving to this dataset indicates that cross-platform variation, not just modality gap, is a major and understudied source of difficulty.
The V2I/I2V asymmetry (visible queries fare much better than infrared queries) implies that thermal imagery loses identity information that visible images retain, motivating work on thermal image enhancement or query-side fusion.
TCC-VPReID's three-stream design gives a reusable template: style augmentation, per-view memory, and an intermediary modality can each be swapped or ablated in future work.

Reading between the lines

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

Aerial infrared may be the main bottleneck: at 15–45 meters altitude with 640×512 resolution, a person's thermal signature is only a few dozen pixels, so the reported 20-point gains may partly reflect how quickly a model learns to exploit clothing-independent thermal silhouettes—an exploitable shortcut rather than identity understanding.
Because identities were collected over five months with consent and facial anonymization, the dataset could enable longitudinal studies of appearance change across days and seasons, which current single-day benchmarks do not capture.
The memory-based cross-view stream suggests a testable extension: using the same dual-memory idea to align other domain splits, such as day versus night or occluded versus unobstructed, without retraining.
A practical consequence the authors do not state: a drone-and-thermal-CCTV setup of this kind could allow nighttime search-and-rescue or security operations to match a ground witness description to an aerial thermal sighting, but only if precision at high ranks is much better than the reported roughly 36% Rank-1.

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.

Desk Editor's Note

read the letter

The dataset is the real contribution here. As far as I can tell, AG-VPReID.VIR is the first video Re-ID benchmark to combine aerial and ground platforms with RGB and IR modalities, and that combination fills a genuine gap for 24-hour surveillance research. The collection effort looks serious: 1,837 identities, 4,861 tracklets, manual annotation with cross-verification, ethics approval, and a sensible train/test split. If the labels hold up, this is a useful resource for the community.

The method, TCC-VPReID, is honestly described as a recombination of known components: style augmentation, memory networks, anaglyph/edge-based intermediate representations. That is fine for a baseline, but the paper sells it as a novel architecture with massive gains, and that is where the problems start.

The stress-test concern is real and it lands. Appendix Section 10 explicitly tunes the loss weights λ1–λ4 on the Aerial→Ground I2V protocol, and Table 3 lists Aerial→Ground as one of the four test protocols. Those same weights are then used for all reported results, including the Ground→Ground numbers in Table 4. So the model has effectively been selected on a test split. The baselines were not given the same test-set tuning opportunity, which means the claimed 20–35 point Rank-1 gaps are not an unbiased estimate of algorithmic merit. This is selection bias, not a conspiracy, but it does undermine the central method-superiority claim.

Other issues are more minor but worth noting: results are single runs with no error bars, so we cannot tell whether the 2–3 point gains on HITSZ-VCM and BUPTCampus are meaningful. The appendix architecture table (Table 8) lists components like graph interaction, domain-adversarial alignment, and temporal memory diffusion that do not appear in the method equations in the main text. And the dataset itself is not yet released, so we cannot independently verify the labels or the tracklet counts.

For a reader, the takeaway is: use this paper for the benchmark, not for the method numbers. The dataset deserves serious referee time, but the authors should be asked to move hyperparameter selection to a validation split, report variance across runs, clarify the architecture table, and release the data and code before publication. I would accept it for peer review with those conditions, but I would not cite the method results as evidence.

Worth a reading group discussion, mostly as a case study on how test-set tuning creeps into benchmark papers.

Referee Report

4 major / 5 minor

Summary. The paper introduces AG-VPReID.VIR, claimed to be the first aerial-ground cross-modality video-based person Re-ID dataset, with 1,837 identities, 4,861 tracklets, and 124,855 frames collected from UAV, CCTV, and wearable cameras in RGB and IR. It also proposes TCC-VPReID, a three-stream architecture combining style-robust feature learning, memory-based cross-view adaptation, and intermediary-guided temporal learning, and reports benchmarks on HITSZ-VCM, BUPTCampus, and AG-VPReID.VIR. The paper claims large performance gains over prior methods, especially on AG-VPReID.VIR; however, the evaluation methodology contains a significant selection-bias problem because the loss hyperparameters are tuned on one of the test protocols.

Significance. If the dataset is released with reliable labels, it is a valuable community resource: it is the first dataset combining aerial and ground platforms with RGB-IR modalities for video Re-ID, and it addresses a real gap in the literature. The paper also provides a broad comparison with many existing methods, ablations of the three streams, qualitative results, and explicit ethics procedures. The method's reported superiority, however, is not established by the current experiments because the model is selected on a test protocol and results are reported from single runs, so the practical significance of the proposed architecture remains uncertain.

major comments (4)

[Sec. 10, Table 9; Sec. 5.2; Table 4] The loss hyperparameters λ1–λ4 are selected by maximizing performance on the Aerial→Ground I2V protocol (Table 9), which is explicitly a test protocol in Table 3. The chosen values (λ1=1.0, λ2=1.5, λ3=1.0, λ4=1.5) are then fixed in Sec. 5.2 and used for all reported results, including the headline Ground↔Ground numbers in Table 4. This means the model has been tuned to a test split, while the compared baselines were not given the same tuning opportunity. The reported 20+ point gains therefore conflate algorithmic merit with test-set overfitting and are not unbiased estimates of generalization. The authors must retune hyperparameters on a held-out validation set disjoint from all four test protocols, or explicitly justify why Aerial→Ground I2V is not part of the evaluation used to report final results.
[Sec. 5.3, Table 4] All results are reported from single runs with no standard deviations, confidence intervals, or multiple random seeds. The central claim is a performance gap of more than 20 Rank-1 points over prior methods; without variance estimates, it is impossible to assess whether this gap is stable or an artifact of a particular initialization or data split. The authors should report mean and standard deviation over at least three random seeds for their method and ideally for the main baselines.
[Appendix Table 8 vs. Sec. 4.1–4.4] Table 8 lists architectural components that do not appear in the method equations, including Cross-View and Cross-Modal Graph Interaction, Domain-Adversarial Alignment, Temporal Memory Diffusion, Modality Decoupling, Multi-level Feature Integration, and Cross-stream Knowledge Transfer. None of these components is defined in Sec. 4, and the ablations in Tables 5 and 6 only vary St1, St2, and St3. It is therefore unclear which components are actually active in the model that produces the results in Table 4. The authors should either provide precise definitions and equations for all components claimed to be part of TCC-VPReID, or remove the unimplemented terms from Table 8 so that the described architecture matches the evaluated architecture.
[Sec. 3.2] The dataset labels are the load-bearing foundation of every benchmark number, but the labeling description relies on 'manual review' and 'cross-verification' without any quantitative evidence of label quality. No inter-annotator agreement is reported, no protocol for resolving disagreements is described, and no independent verification subset is provided. If any tracklet mixes identities or if RGB and IR labels are misaligned, all reported results on AG-VPReID.VIR are affected. The authors should report annotation consistency statistics, release an independently labeled subset, or otherwise demonstrate label reliability across modalities and platforms.

minor comments (5)

[Sec. 5.3] The text states that V2I improvements over the best competitors are '33.69–35.11% in Rank-1', but Table 4 shows differences of 33.44, 33.69, and 34.51 against AuxNet, CST, and SAADG respectively (or 35.12 against SGIEL); the range should be corrected.
[Eq. (9)] The 'anaglyph representation' is described as the output of edge detection with an offset k, but k is never defined and the term 'anaglyph' usually refers to a color-channel composite, not an edge-filter output. Please clarify the operator A, the offset, and why this representation is called anaglyph.
[Sec. 3.1 and Table 1] There are frequent spacing inconsistencies such as 'UA V' instead of 'UAV', and Table 1 contains the typo 'HITZS-VCM' instead of 'HITSZ-VCM'. These should be cleaned up throughout.
[Sec. 4.2, Eq. (7)] The platform-specific prompt P^m_{y_i} and the 'dual-branch attention-based decoder' are not defined in enough detail to be implemented; please specify the form of P, how it is generated, and how Eq. (7) is trained.
[Table 3] The footnote states that 1,184 distractor identities are added to the gallery for all I2V experiments, but the rows for I2V do not show these distractors in the tracklet/frame counts. Please clarify whether the reported Q/G counts include or exclude the distractor set.

Circularity Check

1 steps flagged · score 6.0 of 10

Loss hyperparameters are selected on the Aerial→Ground I2V test protocol, and the same protocol's score is then reported as evaluation, making one central performance claim partially circular.

fitted input called prediction [Supplementary Sec. 10 (Hyperparameter Ablation), Table 9; Sec. 5.2; Table 5]
"Supplementary Sec. 10: 'We conducted ablation studies to determine the optimal values for the loss function hyperparameters (λ1, λ2, λ3, and λ4). Tab. 9 shows the performance of our model with different hyperparameter configurations on the Aerial→Ground I2V protocol...' Sec. 5.2: 'Model trains for 120 epochs with hyperparameters λ1 = 1.0, λ2 = 1.5, λ3 = 1.0, and λ4 = 1.5 (see Appendix Sec. 10).'"

Aerial→Ground I2V is one of the reported evaluation protocols (Table 3). The paper explicitly uses this protocol in Table 9 to choose λ1–λ4 by maximizing Rank-1/mAP. The selected configuration (1.0, 1.5, 1.0, 1.5) is then fixed in Sec. 5.2 and used for all results, and Table 5 reports St123's Aerial→Ground I2V score as 19.83/22.61—exactly the values of the selected row in Table 9. Thus the reported score is not an independent generalization estimate; it is the criterion used to fit the loss weights, i.e., a fitted input presented as a prediction.

full rationale

The only exhibited reduction is the hyperparameter search in Supplementary Sec. 10: λ1–λ4 are selected by maximizing Aerial→Ground I2V scores, fixed in Sec. 5.2, and then the same protocol's St123 scores appear in Table 5 as 19.83/22.61. That specific result is the fitting objective, so this portion of the AG-VPReID evaluation is partially circular. I found no load-bearing self-citation chain: references [24]–[26] are the authors' own prior aerial-ground RGB works, but they are used for inspiration and comparison, not to justify the present claims. The external benchmarks (HITSZ-VCM, BUPTCampus) and the new dataset itself provide independent empirical content, so the circularity is confined to the tuned evaluation pathway rather than the whole paper. A score of 6 reflects one fitted-input-called-prediction step affecting a central reported performance claim, without collapsing the entire derivation by definition.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

The central claims rest on dataset-label quality, the train/test split, and two modeling assumptions about style and edge-based modality bridging. The free parameters are dominated by loss weights selected on a test protocol, which weakens the empirical claims.

free parameters (3)

Loss weights lambda1, lambda2, lambda3, lambda4 = 1.0, 1.5, 1.0, 1.5
Selected by hyperparameter search using the Aerial to Ground I2V test protocol (Appendix Sec. 10, Table 9), which is later reported as a headline result. This is fitting hyperparameters on a test set.
Adaptive stream fusion weights = Not reported
Section 4.4 states that adaptive weighting is applied to each stream's contribution based on validation performance, but the weights and validation protocol are not specified.
Edge filter offset k in Eq. (9) = Not specified
The anaglyph or edge representation in Stream 3 depends on an unspecified offset value k, which is neither learned nor given.

assumptions (4)

domain assumption Manually created identity labels are correct and consistent across RGB/IR modalities and across aerial, CCTV, and wearable views.
Section 3.2 asserts that annotators reviewed YOLO tracks and cross-verified labels; all benchmark metrics depend on this being true.
domain assumption A strict separation between training and testing identities is maintained.
Section 5.1 states a strict separation, but the identity counts across protocols in Table 3 are not clearly reconciled, leaving the assumption unverified from the paper.
ad hoc to paper Edge-filtered 'anaglyph' representations preserve identity-discriminative information while removing modality-specific appearance.
Eq. (9) uses a generic edge operator, and Eq. (10) assumes cross-reconstruction between visible and infrared anaglyph features is a valid way to learn modality-invariant identity features. No independent validation is provided.
ad hoc to paper The style-attack features in Eq. (3) retain identity information while removing style information.
The LSA loss in Eq. (4) relies on this assumption, but the paper provides no analysis or independent check that identity is preserved after the normalization-style attack.

how reviews work

0 comments

Cite this review

Pith. "Pith review of AG-VPReID.VIR: Bridging Aerial and Ground Platforms for Video-based Visible-Infrared Person Re-ID." pith.science (2026). https://pith.science/paper/6PKJCYDE

@misc{pith2026250717995,
  author       = {Pith},
  title        = {Pith review of: AG-VPReID.VIR: Bridging Aerial and Ground Platforms for Video-based Visible-Infrared Person Re-ID},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/6PKJCYDE}},
  note         = {Machine review of arXiv:2507.17995}
}

read the original abstract

Person re-identification (Re-ID) across visible and infrared modalities is crucial for 24-hour surveillance systems, but existing datasets primarily focus on ground-level perspectives. While ground-based IR systems offer nighttime capabilities, they suffer from occlusions, limited coverage, and vulnerability to obstructions--problems that aerial perspectives uniquely solve. To address these limitations, we introduce AG-VPReID.VIR, the first aerial-ground cross-modality video-based person Re-ID dataset. This dataset captures 1,837 identities across 4,861 tracklets (124,855 frames) using both UAV-mounted and fixed CCTV cameras in RGB and infrared modalities. AG-VPReID.VIR presents unique challenges including cross-viewpoint variations, modality discrepancies, and temporal dynamics. Additionally, we propose TCC-VPReID, a novel three-stream architecture designed to address the joint challenges of cross-platform and cross-modality person Re-ID. Our approach bridges the domain gaps between aerial-ground perspectives and RGB-IR modalities, through style-robust feature learning, memory-based cross-view adaptation, and intermediary-guided temporal modeling. Experiments show that AG-VPReID.VIR presents distinctive challenges compared to existing datasets, with our TCC-VPReID framework achieving significant performance gains across multiple evaluation protocols. Dataset and code are available at https://github.com/agvpreid25/AG-VPReID.VIR.

Figures

Figures reproduced from arXiv: 2507.17995 by the authors.

**Figure 2.** Our visible-infrared images vs. existing datasets’ visible [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗

**Figure 3.** Challenges in our proposed dataset. comprising 10 distinct sessions. During each session, UAV flights were performed at multiple altitudes (15m, 25m, and 45m), with each flight lasting 15 minutes to ensure comprehensive coverage. We gathered a total of 124,855 images of 1,837 unique identities. The AG-VPReID.VIR dataset includes both RGB and IR images captured across different camera modalities and perspectives. T… view at source ↗

Figures from the paper (4 more)

**Figure 4.** Figure 4: Overview of our proposed three-stream architecture: (1) Style-Robust Stream for handling appearance variations via style aug [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]

**Figure 5.** Figure 5: Visualization of top-20 ranking results on AG-VPReID.VIR dataset under aerial-ground cross-platform settings for I2V and V2I [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]

**Figure 7.** Figure 7: Distribution of identity presence across different cam [PITH_FULL_IMAGE:figures/full_fig_p011_7.png]

**Figure 6.** Figure 6: Data collection map. V denotes visible, IR means in [PITH_FULL_IMAGE:figures/full_fig_p011_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

55 extracted references · 54 canonical work pages

[1]

G. C. Bertocco, F. Andal ´o, and A. Rocha. Unsupervised and self-adaptative techniques for cross-domain person re- identification. IEEE Transactions on Information Forensics and Security, 16:4419–4434, 2021

work page 2021
[2]

C. Chen, M. Ye, M. Qi, J. Wu, J. Jiang, and C.-W. Lin. Structure-aware positional transformer for visible-infrared person re-identification. IEEE Transactions on Image Pro- cessing, 31:2352–2364, 2022

work page 2022
[3]

Chung, K

D. Chung, K. Tahboub, and E. J. Delp. A two stream siamese convolutional neural network for person re-identification. In ICCV, pages 1983–1991, 2017

work page 1983
[4]

P. Dai, R. Ji, H. Wang, Q. Wu, and Y . Huang. Cross-modality person re-identification with generative adversarial training. In IJCAI, volume 1, page 6, 2018

work page 2018
[5]

Dreuw and ORB-HD

M. Dreuw and ORB-HD. deface: Video anonymization by face detection, 2023. Python package version 1.5.0

work page 2023
[6]

Y . Du, C. Lei, Z. Zhao, Y . Dong, and F. Su. Video-based visible-infrared person re-identification with auxiliary sam- ples. IEEE Transactions on Information Forensics and Secu- rity, 19:1313–1325, 2023

work page 2023
[7]

C. Eom, G. Lee, J. Lee, and B. Ham. Video-based person re- identification with spatial and temporal memory networks. In ICCV, pages 12036–12045, 2021

work page 2021
[8]

J. Feng, A. Wu, and W.-S. Zheng. Shape-erased feature learning for visible-infrared person re-identification. In Pro- ceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 22752–22761, 2023

work page 2023

Show all 55 references

[9]

Y . Feng, F. Chen, J. Yu, Y . Ji, F. Wu, T. Liu, S. Liu, X.-Y . Jing, and J. Luo. Cross-modality spatial-temporal transformer for video-based visible-infrared person re- identification. IEEE Transactions on Multimedia, 26:6582– 6594, 2024

2024
[10]

X. Gu, H. Chang, B. Ma, H. Zhang, and X. Chen. Appearance-preserving 3d convolution for video-based per- son re-identification. In ECCV, pages 228–243. Springer, 2020

2020
[11]

K. He, X. Zhang, S. Ren, and J. Sun. Deep residual learn- ing for image recognition. In Proceedings of the IEEE con- ference on computer vision and pattern recognition , pages 770–778, 2016

2016
[12]

R. Hou, H. Chang, B. Ma, S. Shan, and X. Chen. Temporal complementary learning for video person re-identification. In Computer Vision–ECCV 2020: 16th European Confer- ence, Glasgow, UK, August 23–28, 2020, Proceedings, Part XXV 16, pages 388–405. Springer, 2020

2020
[13]

Jocher, Ayush, and J

G. Jocher, Ayush, and J. Qiu. Ultralytics YOLO. https:// github . com / ultralytics / ultralytics, 2024. Accessed: 2024-03-22

2024
[14]

H. Li, M. Liu, Z. Hu, F. Nie, and Z. Yu. Intermediary- guided bidirectional spatial–temporal aggregation network for video-based visible-infrared person re-identification. IEEE Transactions on Circuits and Systems for Video Tech- nology, 33(9):4962–4972, 2023

2023
[15]

J. Li, S. Zhang, and T. Huang. Multi-scale 3d convolution network for video based person re-identification. In AAAI, volume 33, pages 8618–8625, 2019

2019
[16]

Z. Li, W. Liu, X. Chang, L. Yao, M. Prakash, and H. Zhang. Domain-aware unsupervised cross-dataset per- son re-identification. In Advanced Data Mining and Appli- cations, 2019

2019
[17]

S. Liao, Y . Hu, X. Zhu, and S. Z. Li. Person re-identification by local maximal occurrence representation and metric learning. In CVPR, pages 2197–2206, 2015

2015
[18]

X. Lin, J. Li, Z. Ma, H. Li, S. Li, K. Xu, G. Lu, and D. Zhang. Learning modal-invariant and temporal-memory for video- based visible-infrared person re-identification. In Proceed- ings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 20973–20982, 2022

2022
[19]

H. Liu, X. Tan, and X. Zhou. Parameter sharing exploration and hetero-center triplet loss for visible-thermal person re- identification. IEEE Transactions on Multimedia, 23:4414– 4425, 2020

2020
[20]

W. Liu, X. Chang, L. Chen, D. Q. Phung, X. Zhang, Y . Yang, and A. G. Hauptmann. Pair-based uncertainty and diversity promoting early active learning for person re-identification. ACM Transactions on Intelligent Systems and Technology (TIST), 11:1–15, 2020

2020
[21]

Y . Liu, Z. Yuan, W. Zhou, and H. Li. Spatial and temporal mutual promotion for video-based person re-identification. In AAAI, volume 33, pages 8786–8793, 2019

2019
[22]

McLaughlin, J

N. McLaughlin, J. M. Del Rincon, and P. Miller. Re- current convolutional network for video-based person re- identification. In CVPR, pages 1325–1334, 2016

2016
[23]

D. T. Nguyen, H. G. Hong, K. W. Kim, and K. R. Park. Person recognition system based on a combination of body images from visible light and thermal cameras. Sensors, 17(3):605, 2017

2017
[24]

Nguyen, K

H. Nguyen, K. Nguyen, S. Sridharan, and C. Fookes. Aerial- ground person re-id. In 2023 IEEE International Conference on Multimedia and Expo (ICME) , pages 2585–2590. IEEE, 2023

2023
[25]

Nguyen, K

H. Nguyen, K. Nguyen, S. Sridharan, and C. Fookes. Ag- reid. v2: Bridging aerial and ground views for person re- identification. IEEE Transactions on Information Forensics and Security, 2024

2024
[26]

Nguyen, C

K. Nguyen, C. Fookes, S. Sridharan, F. Liu, X. Liu, A. Ross, D. Michalski, H. Nguyen, D. Deb, M. Kothari, et al. Ag- reid 2023: Aerial-ground person re-identification challenge results. In 2023 IEEE International Joint Conference on Bio- metrics (IJCB), pages 1–10. IEEE, 2023

2023
[27]

H. Park, S. Lee, J. Lee, and B. Ham. Learning by aligning: Visible-infrared person re-identification using cross-modal correspondences. In Proceedings of the IEEE/CVF interna- tional conference on computer vision , pages 12046–12055, 2021

2021
[28]

Radford, J

A. Radford, J. W. Kim, C. Hallacy, A. Ramesh, G. Goh, S. Agarwal, G. Sastry, A. Askell, P. Mishkin, J. Clark, et al. Learning transferable visual models from natural language supervision. In International Conference on Machine Learn- ing, pages 8748–8763. PMLR, 2021

2021
[29]

X. Tian, Z. Zhang, S. Lin, Y . Qu, Y . Xie, and L. Ma. Farewell to mutual information: Variational distillation for cross-modal person re-identification. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 1522–1531, 2021

2021
[30]

Vaswani, N

A. Vaswani, N. Shazeer, N. Parmar, J. Uszkoreit, L. Jones, A. N. Gomez, Ł. Kaiser, and I. Polosukhin. Attention is all you need. In Advances in Neural Information Processing Systems, pages 5998–6008, 2017

2017
[31]

G. Wang, T. Zhang, J. Cheng, S. Liu, Y . Yang, and Z. Hou. Rgb-infrared cross-modality person re-identification via joint pixel and feature alignment. In ICCV, pages 3623– 3632, 2019

2019
[32]

Z. Wang, Z. Wang, Y . Wu, J. Wang, and S. Satoh. Beyond intra-modality discrepancy: A comprehensive survey of het- erogeneous person re-identification. In International Joint Conference on Artificial Intelligence, 2020

2020
[33]

Z. Wang, Z. Wang, Y . Zheng, Y .-Y . Chuang, and S. Satoh. Learning to reduce dual-level discrepancy for infrared- visible person re-identification. In CVPR, pages 618–626, 2019

2019
[34]

Wu, W.-S

A. Wu, W.-S. Zheng, H.-X. Yu, S. Gong, and J. Lai. Rgb- infrared cross-modality person re-identification. In Proceed- ings of the IEEE international conference on computer vi- sion, pages 5380–5389, 2017

2017
[35]

J. Wu, L. He, W. Liu, Y . Yang, Z. Lei, T. Mei, and S. Z. Li. Cavit: Contextual alignment vision transformer for video ob- ject re-identification. In European Conference on Computer Vision, pages 549–566. Springer, 2022

2022
[36]

Q. Wu, P. Dai, J. Chen, C.-W. Lin, Y . Wu, F. Huang, B. Zhong, and R. Ji. Discover cross-modality nuances for visible-infrared person re-identification. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 4330–4339, 2021

2021
[37]

Y . Wu, Y . Lin, X. Dong, Y . Yan, W. Ouyang, and Y . Yang. Exploit the unknown gradually: One-shot video-based per- son re-identification by stepwise learning. In Proceedings of the IEEE conference on computer vision and pattern recog- nition, pages 5177–5186, 2018

2018
[38]

S. Xu, Y . Cheng, K. Gu, Y . Yang, S. Chang, and P. Zhou. Jointly attentive spatial-temporal pooling networks for video-based person re-identification. In ICCV, pages 4733–4742, 2017

2017
[39]

M. Yang, Z. Huang, P. Hu, T. Li, J. Lv, and X. Peng. Learn- ing with twin noisy labels for visible-infrared person re- identification. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 14308– 14317, 2022

2022
[40]

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

2020
[41]

M. Ye, X. Lan, J. Li, and P. Yuen. Hierarchical discriminative learning for visible thermal person re-identification. In Pro- ceedings of the AAAI Conference on Artificial Intelligence , volume 32, 2018

2018
[42]

M. Ye, W. Ruan, B. Du, and M. Z. Shou. Channel aug- mented joint learning for visible-infrared recognition. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 13567–13576, 2021

2021
[43]

M. Ye, J. Shen, D. J. Crandall, L. Shao, and J. Luo. Dynamic dual-attentive aggregation learning for visible-infrared per- son re-identification. In ECCV, pages 229–247. Springer, 2020

2020
[44]

M. Ye, J. Shen, G. Lin, T. Xiang, L. Shao, and S. C. H. Hoi. Deep learning for person re-identification: A survey and out- look. IEEE transactions on pattern analysis and machine intelligence, PP, 2021

2021
[45]

C. Yu, X. Liu, Y . Wang, P. Zhang, and H. Lu. Tf-clip: Learn- ing text-free clip for video-based person re-identification. In Proceedings of the AAAI conference on artificial intelli- gence, volume 38, pages 6764–6772, 2024

2024
[46]

Zhang, H

G. Zhang, H. Zhang, W. Lin, A. K. Chandran, and X. Jing. Camera contrast learning for unsupervised person re-identification. IEEE Transactions on Circuits and Systems for Video Technology, 33(8):4096–4107, 2023

2023
[47]

Zhang, W

S. Zhang, W. Luo, D. Cheng, Q. Yang, L. Ran, Y . Xing, and Y . Zhang. Cross-platform video person reid: A new bench- mark dataset and adaptation approach. In European Confer- ence on Computer Vision (ECCV), 2024

2024
[48]

Zhang and H

Y . Zhang and H. Wang. Diverse embedding expan- sion network and low-light cross-modality benchmark for visible-infrared person re-identification. arXiv preprint arXiv:2303.14481, 2023

2023 arXiv
[49]

Zhang, Y

Y . Zhang, Y . Yan, Y . Lu, and H. Wang. Towards a uni- fied middle modality learning for visible-infrared person re- identification. In Proceedings of the 29th ACM International Conference on Multimedia, pages 788–796, 2021

2021
[50]

Zheng, J

D. Zheng, J. Xiao, M. Sun, H. Bai, and J. Hou. Plausible proxy mining with credibility for unsupervised person re- identification. IEEE Transactions on Circuits and Systems for Video Technology, 33(7):3308–3318, 2022

2022
[51]

C. Zhou, J. Li, H. Li, G. Lu, Y . Xu, and M. Zhang. Video- based visible-infrared person re-identification via style dis- turbance defense and dual interaction. In Proceedings of the 31st ACM International Conference on Multimedia , pages 46–55, 2023. AG-VPReID.VIR: Bridging A...

2023
[52]

6 illustrates the strategic camera placement used for our AG-VPReID.VIR dataset collection across a university campus

Dataset Collection Fig. 6 illustrates the strategic camera placement used for our AG-VPReID.VIR dataset collection across a university campus. The map highlights the diverse sensing infrastruc- ture deployed to capture comprehensive cross-platform and cross-modality data: visi...
[53]

7 illustrates the distribution of data across differ- ent camera types and modalities in our AG-VPReID.VIR dataset

Dataset Distribution Analysis Fig. 7 illustrates the distribution of data across differ- ent camera types and modalities in our AG-VPReID.VIR dataset. The distribution shows a deliberate balance be- tween RGB and IR modalities while maintaining diversity across platforms. UA V...
[54]

The three complementary streams work in concert to overcome the complex variations encountered in cross-platform cross-modality scenarios

Challenges Addressed and Key Components of streams in the proposed architecture Table 8 presents the comprehensive architecture of our TCC-VPReID framework, detailing how each stream ad- dresses specific challenges in aerial-ground visible-infrared person re-identification. Th...
[55]

Hyperparameter Ablation Study We conducted ablation studies to determine the op- timal values for the loss function hyperparameters ( λ1, λ2, λ3, and λ4). Tab. 9 shows the performance of our model with different hyperparameter configurations on the Aerial→Ground I2V protocol, ...

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.

[1] [1]

G. C. Bertocco, F. Andal ´o, and A. Rocha. Unsupervised and self-adaptative techniques for cross-domain person re- identification. IEEE Transactions on Information Forensics and Security, 16:4419–4434, 2021

work page 2021

[2] [2]

C. Chen, M. Ye, M. Qi, J. Wu, J. Jiang, and C.-W. Lin. Structure-aware positional transformer for visible-infrared person re-identification. IEEE Transactions on Image Pro- cessing, 31:2352–2364, 2022

work page 2022

[3] [3]

Chung, K

D. Chung, K. Tahboub, and E. J. Delp. A two stream siamese convolutional neural network for person re-identification. In ICCV, pages 1983–1991, 2017

work page 1983

[4] [4]

P. Dai, R. Ji, H. Wang, Q. Wu, and Y . Huang. Cross-modality person re-identification with generative adversarial training. In IJCAI, volume 1, page 6, 2018

work page 2018

[5] [5]

Dreuw and ORB-HD

M. Dreuw and ORB-HD. deface: Video anonymization by face detection, 2023. Python package version 1.5.0

work page 2023

[6] [6]

Y . Du, C. Lei, Z. Zhao, Y . Dong, and F. Su. Video-based visible-infrared person re-identification with auxiliary sam- ples. IEEE Transactions on Information Forensics and Secu- rity, 19:1313–1325, 2023

work page 2023

[7] [7]

C. Eom, G. Lee, J. Lee, and B. Ham. Video-based person re- identification with spatial and temporal memory networks. In ICCV, pages 12036–12045, 2021

work page 2021

[8] [8]

J. Feng, A. Wu, and W.-S. Zheng. Shape-erased feature learning for visible-infrared person re-identification. In Pro- ceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 22752–22761, 2023

work page 2023