Pith. sign in

REVIEW 3 major objections 5 minor 27 references

Real-time Person Re-identification at the Edge: A Mixed Precision Approach

T0 review · 3 major / 5 minor · reviewed 2026-08-14 · deepseek-v4-flash

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

Pith's one-line read A person re-identification system built on MobileNet-V2 and trained in mixed precision achieves real-time edge inference at 27.77 fps and 6.48 W, with only a 5.6% average rank-1 accuracy loss over a ResNet-50 single-precision baseline.

desk verdict Useful edge-deployment study, but the headline speedup/power numbers conflate architecture, precision, and hardware, and the accuracy claim doesn't match the paper's own figures. read the letter →

arxiv 1908.07842 v1 pith:FMYOAJPC submitted 2019-08-19 cs.LG stat.ML

Mohammadreza Baharani , Shrey Mohan , Hamed Tabkhi This is my paper · ORCID
classification cs.LGstat.ML
keywords personre-identificationmixedprecisiontrainingedgecomputingMobileNet-V2ResNet-50tripletlossNVIDIAXavierreal-timeinference
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 argues that real-time person re-identification—recognizing the same person across different camera views—does not require a heavy server-side network. It shows that a MobileNet-V2 backbone trained with mixed FP32/FP16 precision and deployed in half precision on an NVIDIA Xavier edge node identifies people at 27.77 frames per second while consuming 6.48 W. Against a ResNet-50 single-precision baseline on CUHK03, Market-1501, and DukeMTMC, this is a 3.25x throughput gain, a 1.45x power reduction, and 1.75x faster training, at an average accuracy cost of 5.6%. The paper also pins down which parts of a re-ID network tolerate half precision: convolutions and matrix multiplies can run in FP16, but batch-normalization inputs and the triplet loss must stay in FP32 for training to converge and hard-sample mining to work.

What carries the argument

The central mechanism is mixed-precision training configured by a specific layer partition: convolution and general matrix multiply (GeMM) operations are assigned to half precision (FP16), while batch-normalization inputs and the triplet-loss calculation remain in single precision (FP32). The paper reports that FP16 batch-norm inputs stop training from converging, and an FP16 loss computation degrades the hard-sample pool used for triplet mining. The MobileNet-V2 backbone provides the light-weight representation, the triplet loss with hard mining provides the re-identification objective, and on the NVIDIA Xavier edge node the FP16 models run on the Deep Learning Accelerators while FP32 models run on the Volta GPU cores.

What would settle it

Run the identical MobileNet-V2 model in FP32 and FP16 on the same processor (both on Volta GPU cores, or both on the Deep Learning Accelerator) and measure throughput, power, and accuracy; if FP16 does not improve throughput or cut power, the paper's precision-driven gains collapse.

Watch

Extended reading notes

Core claim

The paper claims that mixed precision is a usable deployment lever for person re-ID at the edge, not just a training-time trick. Concretely, replacing the ResNet-50 backbone with MobileNet-V2 and training in mixed precision lets inference run at 27.77 fps and 6.48 W on NVIDIA Xavier while keeping rank-1 accuracy within 5.6% of the ResNet-50 single-precision baseline on three benchmarks. The authors further report that half precision alone degrades ResNet-50 by only 0.9% on average, and that the FP16 MobileNet-V2 model is 18.92x smaller than the baseline. The central engineering rule that emerges is a layer-wise precision partition: convolution and GeMM operations go to FP16, while batch-normalization inputs and the triplet-loss computation stay in FP32, because FP16 in those places prevents convergence and weakens the hard-positive/negative mining that triplet loss depends on.

Load-bearing premise

The headline speed and power numbers compare two networks running on two different processors at the same time, so they are not a clean measurement of what mixed precision alone buys.

Editorial extensions

If this is right

  • At 27.77 fps, the FP16 MobileNet-V2 model clears the common 25 fps real-time video bar, so a single edge node can keep pace with live camera streams.
  • The precision-partitioning rule—conv and GeMM in FP16, batch-norm inputs and loss in FP32—is presented as transferable to other re-ID networks, allowing existing high-accuracy designs to move to the edge without architectural changes.
  • The model size drops from 94.6 MB for ResNet-50 single precision to 5.0 MB for FP16 MobileNet-V2, easing storage and memory-bandwidth constraints on embedded hardware.
  • Training time falls roughly from 242.65 to 138.1 minutes for the MobileNet-V2 comparison, so mixed precision also shortens the server-side tuning loop.
  • Because half precision alone costs ResNet-50 only about 0.9% average accuracy, precision reduction appears to be a cheaper deployment knob than backbone substitution.

Reading between the lines

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

  • The paper's tables imply a distinction its abstract does not make explicit: the headline 3.25x throughput and 1.45x power gains compare MobileNet-V2 FP16 on the Deep Learning Accelerator against ResNet-50 FP32 on the Volta GPU cores, so the bundled gain includes the backbone swap and the compute-unit change, not just precision.
  • A clean test would hold the compute unit fixed and compare FP32 versus FP16 for the same network; the paper does not report that comparison, so the pure precision contribution to throughput and power remains untested.
  • Because the accuracy cost of half precision for ResNet-50 is much smaller than the cost of switching to MobileNet-V2, most of the measured speedup likely comes from the architecture swap rather than from FP16 arithmetic.
  • The reported 27.77 fps covers feature extraction only, not detection or tracking; an end-to-end multi-camera tracking pipeline would consume a larger share of the edge node's time and power budget.
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 / 5 minor

Summary. This manuscript reports an empirical study of person re-identification at the edge. The authors fine-tune ResNet-50 and MobileNet-V2 with a triplet loss using either single precision or mixed precision, evaluate on CUHK03, Market-1501, and DukeMTMC, and measure training time, model size, inference throughput, and power on an NVIDIA Xavier. They report that MobileNet-V2 with mixed precision reaches 27.77 fps at 6.48 W with an average accuracy loss of 5.6% relative to the ResNet-50 single-precision baseline, and claim 3.25x throughput, 1.75x training-time, and 1.45x power improvements. The paper includes a qualitative comparison and the authors state that code and pre-trained networks are publicly available.

Significance. The paper addresses a relevant deployment problem: real-time person re-ID on constrained hardware. Its strengths are the reproducible artifacts (public code/models), realistic measurements on a current edge platform, and an empirical comparison on three standard benchmarks. If the stated contributions are re-framed carefully, the system-level comparison is a useful engineering data point for practitioners. The claimed methodological finding—that batch-normalization inputs and the triplet loss must remain in FP32 during mixed-precision training—is interesting but currently lacks ablation support. The main limitation is that the headline speed, power, and training-time numbers are presented as effects of mixed precision even though the experiments change the backbone, precision, and compute unit simultaneously.

major comments (3)
  1. [Section 4.4 (Tables 6-7), Section 4.3 (Table 3), Abstract] The headline improvements are not cleanly attributable to mixed precision. Section 4.4 states that half-precision models run on the Xavier DLAs while single-precision models run on the Volta GPU cores. Consequently, Table 6's 'Over the Baseline' figure of 3.25x compares MobileNet-V2 FP16 on DLA with ResNet-50 FP32 on Volta, and Table 7's 1.45x power improvement is the ratio of ResNet-50 FP32 power to MobileNet-V2 FP16 power. Even the 'Per Same Model' columns (1.38x throughput, 1.0x power for MobileNet-V2) are not precision-only comparisons because the mixed-precision rows use DLA and the single-precision rows use Volta. Likewise, the 1.75x training-time speedup in Table 3 is the ratio of ResNet-50 single-precision time (242.65 min) to MobileNet-V2 mixed-precision time (138.1 min), while the same-model speedup for MobileNet-V2 is only 140.3/138.1 = 1.016x. The abstract therefore overstates what is measured. Please either add a same-compute-unit precision comparison (e.g., FP16 vs FP32 on the Volta GPU) or rephrase the claims as system-level gains from changing the backbone, precision, and compute engine together.
  2. [Abstract, Section 4.2, Figure 5] The claim that MobileNet-V2 mixed precision 'deteriorates accuracy only 5.6%' cannot be derived from the reported results. Figure 5 gives mean CMC-1 values of 76.4 for ResNet-50 single precision and 70.3 for MobileNet-V2 mixed precision, a 6.1 percentage-point gap (approximately 8% relative), not 5.6%. No other metric in Figures 3-5 yields 5.6% as an average over the three datasets. Please state the exact metric and computation behind the 5.6% figure or correct the abstract and Section 4.2. In addition, no variance or number of repeats is reported, so the 'negligible' 0.5-point difference between MobileNet-V2 single and mixed precision is not established as meaningful.
  3. [Section 3.3] The key empirical rule that 'inputs of batch normalization layers and the loss calculation should be in FP32' is stated as a finding, but no experiment supports it. The reader cannot tell whether FP16 BN inputs actually prevent convergence, whether FP16 loss harms hard-sample mining, or whether these choices merely have a small effect. Since this is one of the paper's stated contributions and is used to justify the mixed-precision design, please include an ablation or training curves with and without the FP32 constraint.
minor comments (5)
  1. [Section 3.1.2] The citation 'Andrew et al.[4]' should refer to the MobileNets paper (reference [10] in the bibliography), not to the ImageNet reference [4].
  2. [Section 1 and Section 3.1.1] There are typos: 'relive the pressure' should be 'relieve the pressure' in Section 1, and 'concider' should be 'consider' in Section 3.1.1.
  3. [Section 4.2] The sentence 'single precision negligibility deteriorate the CMC-1 performance for 0.5%' is ungrammatical and seems to state the opposite of the data; please rewrite to say that mixed precision is 0.5 percentage points below single precision.
  4. [Figures 3-5] The in-bar labels in Figures 3-5 are very small, and the exact mean values matter for the paper's accuracy claims; please also provide the mean values in a table.
  5. [Table 5] The column headings of Table 5 are confusing: it is not immediately clear whether 'Per Same Model' and 'Over the Baseline' apply to both precision columns. Please reformat the table so the improvement factors are unambiguous.

Circularity Check

0 steps flagged · score 0.0 of 10

No circular derivation: all results are measured on external benchmarks and independent prior methods; reported gains are experimental measurements, not constructed predictions.

full rationale

The paper does not derive any result from assumptions that already contain the conclusion. Its contributions are empirical: it takes two established backbones (ResNet-50 and MobileNet-V2), trains them in single and mixed precision using NVIDIA Apex, and measures accuracy on the external benchmarks CUHK03, Market-1501, and DukeMTMC, plus training time and edge throughput/power on an NVIDIA Xavier. The headline numbers (3.25x throughput, 1.45x power reduction, 5.6% accuracy deterioration) are direct measurements reported in Tables 3, 6, and 7 and Figure 3-5; no fitted parameter is renamed a prediction, and no quantity is defined in terms of another quantity that is then claimed as an output. Mixed precision training is imported from the independent prior work of Micikevicius et al. [14] and the Apex library, MobileNet-V2 from Sandler et al. [11], and the triplet loss from Hermans et al. [8]; none of these load-bearing elements is a self-citation of the present authors, and no self-citation is used to justify an otherwise unsupported premise. The one legitimate concern is interpretive rather than circular: the abstract's 'Over the Baseline' comparison changes the backbone, the precision, and the compute unit simultaneously, because Section 4.4 states 'We mapped the half-precision types of both networks on Deep Learning Accelerators (DLA) and single precision to GPU Volta cores,' so the 3.25x and 1.45x gains are system-level effects rather than a clean measurement of mixed precision alone; the same tables' 'Per Same Model' columns disclose the precision-only comparison (1.38x throughput, 1.0x power for MobileNet-V2). This is a disclosed experimental-confound issue, not a circular derivation. Likewise, the abstract's 5.6% accuracy gap is not exactly reproduced by Fig. 5's means (76.4 vs 70.3 yields 6.1 percentage points), which is a possible numerical inconsistency but again not circularity. Because all central claims are externally measured and independently sourced, no circular step exists.

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

The paper is an empirical systems evaluation with no mathematical derivation. The claims rest on: (1) transferability of ImageNet-pretrained backbones to re-ID, (2) the triplet loss objective, (3) the Apex mixed precision recipe, and (4) an ad hoc constraint on which layers can be half precision, discovered by the authors. The hand-chosen hyperparameters (margin, learning rate, batch structure) are standard but untested for sensitivity, and the paper provides no independent evidence for them.

free parameters (4)
  • Triplet loss margin = 0.3
    Section 3.2, Table 1: margin is set to 0.3 without sensitivity analysis; accuracy depends on this choice.
  • Initial learning rate = 2e-4
    Table 1 sets Adam initial LR to 2e-4; no ablation or schedule details beyond exponential decay after epoch 150.
  • Learning rate decay = unspecified exponential rate
    Section 4.1 says LR decreases exponentially after 150 epochs but the rate is not given; results depend on it.
  • Batch composition = 32 IDs, 4 instances per ID
    Table 1; triplet sampling structure, chosen by hand, standard but untested here.
assumptions (4)
  • domain assumption ImageNet-pretrained backbones transfer to person re-ID.
    Section 3.1 uses ImageNet-pretrained ResNet-50 and MobileNet-V2 with the final layer replaced; transferability is assumed.
  • domain assumption Triplet loss with hard sample mining is an effective re-ID objective.
    Section 3.2 adopts the triplet loss of Hermans et al. [8] without comparing to other losses.
  • standard math Apex mixed-precision training (FP32 master weights, loss scaling) preserves convergence.
    Section 3.3 invokes Micikevicius et al. [14]; the paper contributes no new theory here.
  • ad hoc to paper Batch norm inputs and loss must remain FP32 for re-ID mixed-precision training to converge.
    Section 3.3 reports this as an experimental finding; it is a constraint unique to this paper's setup.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Real-time Person Re-identification at the Edge: A Mixed Precision Approach." pith.science (2026). https://pith.science/paper/FMYOAJPC

@misc{pith2026190807842,
  author       = {Pith},
  title        = {Pith review of: Real-time Person Re-identification at the Edge: A Mixed Precision Approach},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/FMYOAJPC}},
  note         = {Machine review of arXiv:1908.07842}
}
abstract

A critical part of multi-person multi-camera tracking is person re-identification (re-ID) algorithm, which recognizes and retains identities of all detected unknown people throughout the video stream. Many re-ID algorithms today exemplify state of the art results, but not much work has been done to explore the deployment of such algorithms for computation and power constrained real-time scenarios. In this paper, we study the effect of using a light-weight model, MobileNet-v2 for re-ID and investigate the impact of single (FP32) precision versus half (FP16) precision for training on the server and inference on the edge nodes. We further compare the results with the baseline model which uses ResNet-50 on state of the art benchmarks including CUHK03, Market-1501, and Duke-MTMC. The MobileNet-V2 mixed precision training method can improve both inference throughput on the edge node, and training time on server $3.25\times$ reaching to 27.77fps and $1.75\times$, respectively and decreases power consumption on the edge node by $1.45\times$, while it deteriorates accuracy only 5.6\% in respect to ResNet-50 single precision on the average for three different datasets. The code and pre-trained networks are publicly available at https://github.com/TeCSAR-UNCC/person-reid.

Figures

Figures reproduced from arXiv: 1908.07842 by the authors.

Figure 1
Figure 1. Structural components of ResNet 3 Mixed Precision Real-time Person Re-identification In this section, we discuss two ResNet-50 and MobileNet-V2 architectures. Then we continue to explain the loss function, and in the last, we give an introduction about mixed precision training, and we elaborate more on network layers and loss precision partitioning. 3.1 Background In this section, we will give an introduction for tw… view at source ↗
Figure 2
Figure 2. Different MobileNet-V2 convolution blocks [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. ResNet-50 accuracy evaluation on three different benchmarks. We trained the model for two different [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: MobileNetV2 accuracy evaluation on three different benchmarks. We trained the model for two different [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]
Figure 5
Figure 5. Figure 5: A CMC-1 Comparison of two networks ResNet50 (RN) and MobileNetV2 (MN) for three different training [PITH_FULL_IMAGE:figures/full_fig_p006_5.png]
Figure 6
Figure 6. Figure 6: A qualitative comparison of two networks with different precision. The images without the bounding box in [PITH_FULL_IMAGE:figures/full_fig_p007_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

27 extracted references · 22 canonical work pages

  1. [1]

    Baharani, M., Noori, H., Aliasgari, M., Navabi, Z.: High-level design space exploration of locally linear neuro-fuzzy models for embedded systems. Fuzzy Sets and Systems 253, 44–63 (2014) 5https://developer.nvidia.com/embedded/jetson-agx-xavier-dl-inference-benchmarks 8 PUBLISHED AS A CONFERENCE PAPER AT ICIAR 2019 - A UGUST 22, 2019

    work page 2014

  2. [2]

    In: 2011 8th IEEE International Conference on Advanced Video and Signal Based Surveillance (A VSS)

    Bk, S., Corvee, E., Brmond, F., Thonnat, M.: Multiple-shot human re-identification by mean riemannian covariance grid. In: 2011 8th IEEE International Conference on Advanced Video and Signal Based Surveillance (A VSS). pp. 179–184 (2011)

    work page 2011

  3. [3]

    In: 2009 Fourth International Conference on Embedded and Multimedia Computing

    de Oliveira, I.O., de Sousa Pio, J.L.: Object reidentification in multiple cameras system. In: 2009 Fourth International Conference on Embedded and Multimedia Computing. pp. 1–8 (2009)

    work page 2009

  4. [4]

    In: 2009 IEEE Conference on Computer Vision and Pattern Recognition

    Deng, J., Dong, W., Socher, R., Li, L., and: Imagenet: A large-scale hierarchical image database. In: 2009 IEEE Conference on Computer Vision and Pattern Recognition. pp. 248–255 (2009)

    work page 2009

  5. [5]

    Person Re-Identification pp

    Fleuret, F., Ben Shitrit, H., Fua, P.: Re-identification for improved people tracking. Person Re-Identification pp. 309–330 (2014)

    work page 2014

  6. [6]

    In: Proceedings of the Association for the Advancement of Artificial Intelligence (AAAI) (2019)

    Fu, Y ., Wei, Y ., Zhou, Y ., Shi, H., Huang, G., Wang, X., Yao, Z., Huang, T.S.: Horizontal pyramid matching for person re-identification. In: Proceedings of the Association for the Advancement of Artificial Intelligence (AAAI) (2019)

    work page 2019

  7. [7]

    In: 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)

    He, K., Zhang, X., Ren, S., Sun, J.: Deep residual learning for image recognition. In: 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). pp. 770–778 (2016)

    2016

  8. [8]

    CoRR abs/1703.07737 (2017)

    Hermans, A., Beyer, L., Leibe, B.: In defense of the triplet loss for person re-identification. CoRR abs/1703.07737 (2017)

    arXiv 2017

Show all 27 references
  1. [9]

    In: Heyden, A., Kahl, F

    Hirzer, M., Beleznai, C., Roth, P.M., Bischof, H.: Person re-identification by descriptive and discriminative classification. In: Heyden, A., Kahl, F. (eds.) Image Analysis. pp. 91–102. Springer Berlin Heidelberg, Berlin, Heidelberg (2011)

    2011

  2. [10]

    CoRR abs/1704.04861 (2017)

    Howard, A.G., Zhu, M., Chen, B., Kalenichenko, D., Wang, W., Weyand, T., Andreetto, M., Adam, H.: Mobilenets: Efficient convolutional neural networks for mobile vision applications. CoRR abs/1704.04861 (2017)

    2017 arXiv

  3. [11]

    In: 2018 IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2018, Salt Lake City, UT, USA, June 18-22, 2018

    Jacob, B., Kligys, S., Chen, B., Zhu, M., Tang, M., Howard, A.G., Adam, H., Kalenichenko, D.: Quantization and training of neural networks for efficient integer- arithmetic-only inference. In: 2018 IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2018, Salt Lake...

    2018

  4. [12]

    In: 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition

    Li, S., Bak, S., Carr, P., Wang, X.: Diversity regularized spatiotemporal attention for video-based person re-identification. In: 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. pp. 369–378 (2018)

    2018

  5. [13]

    In: CVPR (2014)

    Li, W., Zhao, R., Xiao, T., Wang, X.: Deepreid: Deep filter pairing neural network for person re-identification. In: CVPR (2014)

    2014

  6. [14]

    In: International Conference on Learning Representations (2018)

    Micikevicius, P., Narang, S., Alben, J., Diamos, G., Elsen, E., Garcia, D., Ginsburg, B., Houston, M., Kuchaiev, O., Venkatesh, G., Wu, H.: Mixed precision training. In: International Conference on Learning Representations (2018)

    2018

  7. [15]

    In: 2013 Conference on Design and Architectures for Signal and Image Processing

    Mier y Tern, A.R., Lacassagne, L., Zahraee, A.H., Gouiffs, M.: Real-time covariance tracking algorithm for embedded systems. In: 2013 Conference on Design and Architectures for Signal and Image Processing. pp. 104–111 (2013)

    2013

  8. [16]

    In: European Conference on Computer Vision workshop on Benchmarking Multi-Target Tracking (2016)

    Ristani, E., Solera, F., Zou, R., Cucchiara, R., Tomasi, C.: Performance measures and a data set for multi-target, multi-camera tracking. In: European Conference on Computer Vision workshop on Benchmarking Multi-Target Tracking (2016)

    2016

  9. [17]

    In: Conference on Computer Vision and Pattern Recognition (2018)

    Ristani, E., Tomasi, C.: Features for multi-target multi-camera tracking and re-identification. In: Conference on Computer Vision and Pattern Recognition (2018)

    2018

  10. [18]

    In: The European Conference on Computer Vision (ECCV) (September 2018)

    Shen, Y ., Li, H., Yi, S., Chen, D., Wang, X.: Person re-identification with deep similarity-guided graph neural network. In: The European Conference on Computer Vision (ECCV) (September 2018)

    2018

  11. [19]

    In: International Conference on Learning Representations (2015)

    Simonyan, K., Zisserman, A.: Very deep convolutional networks for large-scale image recognition. In: International Conference on Learning Representations (2015)

    2015

  12. [20]

    In: Proceedings of the European Conference on Computer Vision (ECCV)

    Sun, Y ., Zheng, L., Yang, Y ., Tian, Q., Wang, S.: Beyond part models: Person retrieval with refined part pooling (and a strong convolutional baseline). In: Proceedings of the European Conference on Computer Vision (ECCV). pp. 480–496 (2018)

    2018

  13. [21]

    In: ACCV Workshops (2012)

    y Ter ´an, A.R.M., Gouiff `es, M., Lacassagne, L.: Covariance descriptor multiple object tracking and re-identification with colorspace evaluation. In: ACCV Workshops (2012)

    2012

  14. [22]

    In: 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)

    Xiao, T., Li, S., Wang, B., Lin, L., Wang, X.: Joint detection and identification feature learning for person search. In: 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). pp. 3376–3385 (2017)

    2017

  15. [23]

    arXiv preprint arXiv:1711.08184 (2017)

    Zhang, X., Luo, H., Fan, X., Xiang, W., Sun, Y ., Xiao, Q., Jiang, W., Zhang, C., Sun, J.: Alignedreid: Surpassing human-level performance in person re- identification. arXiv preprint arXiv:1711.08184 (2017)

    2017 arXiv

  16. [24]

    In: Computer Vision, IEEE International Conference on (2015)

    Zheng, L., Shen, L., Tian, L., Wang, S., Wang, J., Tian, Q.: Scalable person re-identification: A benchmark. In: Computer Vision, IEEE International Conference on (2015)

    2015

  17. [25]

    In: Proceedings of the IEEE International Conference on Computer Vision (2017)

    Zheng, Z., Zheng, L., Yang, Y .: Unlabeled samples generated by gan improve the person re-identification baseline in vitro. In: Proceedings of the IEEE International Conference on Computer Vision (2017)

    2017

  18. [26]

    Zhong, Z., Zheng, L., Cao, D., Li, S.: Re-ranking person re-identification with k-reciprocal encoding (2017)

    2017

  19. [27]

    IEEE Transactions on Image Processing 27(11), 5683–5695 (2018) 9

    Zhu, X., Jing, X., You, X., Zhang, X., Zhang, T.: Video-based person re-identification by simultaneously learning intra-video and inter-video distance metrics. IEEE Transactions on Image Processing 27(11), 5683–5695 (2018) 9

    2018

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.