Multi-Label Classification Framework for Hurricane Damage Assessment

John E. Taylor; Neda Mohammadi; Zhangding Liu

REVIEW 4 major objections 5 minor 1 cited by

Multi-Label Classification Framework for Hurricane Damage Assessment

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 claims that pairing ResNet-152 with the CSRA attention module reaches 90.23% mean average precision for multi-label hurricane damage assessment on the RescueNet dataset, beating plain ResNet-152's 88.19% and other baselines.

desk verdict A plausible two-point mAP gain from adding CSRA to ResNet on RescueNet, but the missing label-conversion details make the headline number unverifiable as written. read the letter →

arxiv 2507.02265 v1 pith:NVRBRXGA submitted 2025-07-03 cs.CV cs.AI

Zhangding Liu , Neda Mohammadi , John E. Taylor This is my paper · ORCID

classification cs.CVcs.AI

keywords multi-labelclassificationhurricanedamageassessmentclass-specificresidualattentionResNetRescueaerialimagerymeanaverageprecision

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 tries to establish that post-hurricane damage assessment should be treated as a multi-label classification problem, since one aerial image can contain several damage types and severities at once, and that a standard convolutional backbone with a class-specific attention module is enough to handle that complexity. The proposed framework pairs a ResNet feature extractor with the CSRA residual attention module and reports a mean average precision of 90.23% on the RescueNet dataset of Hurricane Michael imagery, outperforming plain ResNet-152's 88.19% and other baselines. If the result holds, disaster responders would have a relatively simple recipe: use a pretrained ResNet, add CSRA, train with per-class sigmoid outputs, and get a more complete damage map from each image than single-label systems provide.

What carries the argument

The central object is the class-specific residual attention (CSRA) module mounted on a ResNet backbone. For each class, CSRA computes a spatial attention score that highlights the image regions most relevant to that class, combines that score with the class-agnostic global average pooling feature, and applies residual attention to produce logits that are fused into the final prediction. This design lets the model attend to class-specific locations while keeping the global context, and the sigmoid activation makes the per-class decisions independent instead of competitive.

What would settle it

Re-train the ResNet-152 and ResNet-152+CSRA models with a documented rule for converting RescueNet pixel masks into image-level labels, such as declaring a class present when at least 1%, 5%, or 10% of pixels belong to it, and compare mAP across those settings; if the 90.23% versus 88.19% gap shrinks or reverses for some plausible threshold, the claimed gain comes from the label conversion rather than from CSRA.

Watch

Extended reading notes

Core claim

The central discovery is that adding CSRA to a ResNet backbone yields consistent gains for multi-label hurricane damage classification, with ResNet-152+CSRA reaching 90.23% mAP, 87.37% overall precision, and 84.40% overall F1 on RescueNet. The paper attributes the improvement to CSRA's per-class spatial attention scores, which highlight image regions relevant to each damage category and are combined with class-agnostic global average pooling features; a sigmoid output layer then lets each class be predicted independently. The same pattern appears across depths, with CSRA raising ResNet-50 from 86.55% to 89.64% mAP and ResNet-101 from 87.24% to 89.81% mAP, indicating that the attention module rather than simply network depth is doing the work.

Load-bearing premise

The whole comparison rests on how the pixel-level damage masks in RescueNet were turned into the image-level labels the model trains on and is scored against, and the paper does not describe that conversion.

Editorial extensions

If this is right

One model pass per aerial image can simultaneously label buildings, roads, trees, water, vehicles, and damage levels, replacing pipelines that detect each class separately.
Because predictions are per-class and not mutually exclusive, images with mixed damage, such as a building with both major damage and a blocked road, can be represented honestly in a single label vector.
The roughly two-point mAP gain from adding CSRA suggests that spatial attention over class-relevant regions is a cheap, effective upgrade over plain average pooling for disaster imagery.
On the RescueNet benchmark, CSRA improves every tested ResNet depth, and the best absolute result comes from the largest backbone, ResNet-152.

Reading between the lines

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

The reported mAP depends on how RescueNet's pixel-level semantic labels were converted into image-level multi-labels, and the paper does not document that conversion; different plausible rules could shift the absolute numbers and possibly the ranking between models.
The paper reports only aggregate metrics such as mAP and OP, and its own Discussion says per-class accuracy is left to future work, so class-level failure modes such as rare damage categories are still unmeasured.
A direct extension would be to vary the 0.5 prediction threshold or use adaptive thresholds on imbalanced damage classes, and to check whether the CSRA gain persists on storms other than Hurricane Michael.

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 genuinely new thing here is narrow: this is the first time CSRA has been bolted onto a ResNet backbone for multi-label hurricane damage classification on RescueNet, and they report a specific number (90.23% mAP) not in the prior literature. That is a legitimate applied result, not a new method. The paper is also clearly written, uses a public dataset, and documents the training setup well enough that the comparisons among backbones are probably reproducible.

The soft spots are real and one is load-bearing. The paper never says how image-level multi-label targets were derived from RescueNet's pixel-level semantic segmentation labels. That step determines the ground truth for both training and evaluation, so different conversion rules (pixel-count thresholds, minimum connected components, morphological filtering) would produce different label distributions and could easily shift the reported mAP by more than the two-point gain attributed to CSRA. This is not a minor documentation gap; the headline result is unverifiable without it. Second, there are no error bars or multiple seeds. A 1.5- to 2-point mAP difference between ResNet152 and ResNet152+CSRA could plausibly be within run-to-run noise, and the paper gives no statistical evidence that the improvement is stable. Third, the comparison set is only backbone variations, plus one ViT+CSRA row. No existing multi-label methods (e.g., Hua et al.'s relation network, or transformer-based multi-label classifiers) are included, so "outperforming existing baseline methods" is an overstatement. Fourth, per-class AP is absent; the authors acknowledge this in the Discussion when they say future work will evaluate per-class accuracy, so at least the limitation is on record.

I want to be fair: the central claim is plausible, and the CSRA gain is consistent with the original CSRA paper's results on other multi-label benchmarks. I do not think the numbers are fabricated. But the missing label-generation description alone would force any serious reader to treat the result as conditional.

Who is this for? Researchers working on post-disaster damage classification or multi-label aerial image analysis will find it a useful data point, though not a technique they would adopt wholesale. It deserves a serious referee because the application is relevant, the dataset is public, and the gap is fixable. My recommendation: send it to review, but the revision must document label generation, include variance estimates or multiple seeds, add per-class AP, and compare with at least one non-backbone multi-label method.

Referee Report

4 major / 5 minor

Summary. The paper proposes a multi-label classification framework for post-hurricane damage assessment, combining a ResNet feature extractor with a class-specific residual attention (CSRA) module. Using aerial imagery from the RescueNet dataset after Hurricane Michael, the authors report a mean average precision of 90.23% for ResNet-152+CSRA, outperforming plain backbones (ResNet-50/101/152, VGG19, EfficientNet) and a ViT-B16+CSRA baseline. The central claim is that the CSRA module adds about two mAP points to the ResNet backbones, enabling more comprehensive multi-label damage recognition.

Significance. If the central claim is supported, the contribution is a straightforward but potentially practical combination of an existing attention module with a standard backbone for multi-label disaster damage classification. The paper uses a public dataset and a published attention mechanism, and the reported mAP numbers are internally consistent. However, the absolute mAP values and the rank ordering of models depend entirely on the undocumented conversion from RescueNet's pixel-level semantic segmentation to image-level multi-label targets. The comparison set is limited to backbone variations and does not include established multi-label aerial image classification baselines. The paper also omits per-class AP, error bars, and statistical significance tests, so the strength of the claimed improvement cannot be assessed at this stage.

major comments (4)

[Data section, first paragraph] The manuscript never specifies how image-level multi-label ground truth was derived from RescueNet's pixel-level semantic segmentation masks. The Data section states that the study "utilized aerial imagery and labels from the RescueNet dataset" and shows example multi-labels in Figure 3, but no conversion rule is given, such as whether a class is positive if any pixel of that class is present, or whether a minimum area or morphological filtering threshold is applied. This is a load-bearing omission because the reported mAP is computed against these labels; different conversion rules can materially change the label distribution, class imbalance, and the relative difficulty of the task. The paper should document the conversion step and report the number of positive images per class, especially for rare classes like vehicles and pools, before the results can be trusted.
[Results, Table 1 and Evaluation Metrics] Table 1 reports a single run of each model with no error bars, multiple seeds, or statistical significance tests. The differences among the top-performing models are small; for example, ResNet-152+CSRA (90.23 mAP) and ResNet-101+CSRA (89.81 mAP) differ by 0.42 percentage points, and it is not clear whether this difference is reproducible. The manuscript also does not report per-class average precision, so the reader cannot determine whether the mAP gain from CSRA is spread across classes or driven by a few easy categories. These results should be augmented with variance estimates and, at a minimum, per-class AP values.
[Results and Analysis, Table 1] The experimental comparison is limited to different feature extractors with and without the CSRA module. The paper claims the proposed method "outperforms existing baseline models," but the only external baseline is ViT-B16+CSRA, and no comparison is made to established multi-label aerial image classification methods, such as the relation network cited in the introduction, or to the original RescueNet semantic segmentation framework. Without such comparisons, the claim that the proposed framework outperforms existing methods is not supported by the experiments; the table only demonstrates that ResNet+CSRA is better than other backbones with or without CSRA.
[Discussion, second paragraph] The authors acknowledge in the Discussion that "our study mainly demonstrates that MLCSANet achieves high overall performance, as evidenced by mAP and OP metrics" and that "future work will further evaluate per-class accuracy." This explicit admission of the absence of per-class evaluation undermines the broad conclusion that the method effectively identifies diverse damage types, since mAP can be high even when rare damage categories are poorly recognized. The manuscript should either provide per-class results or temper the abstract and conclusion claims to reflect this limitation.

minor comments (5)

[Discussion, second paragraph] The model name "MLCSANet" is introduced in the Discussion but is not defined or used earlier in the manuscript; the framework is previously referred to as ResNet+CSRA. Please introduce the name in the Methodology section if it is meant to be the proposed model.
[Discussion, second paragraph] There are several typos: "effiency" should be "efficiency," and "adaptive thresholding strategiesand" contains a duplicated "and." A careful proofreading pass is needed.
[Introduction, first paragraph] The reference to He et al. is cited as 2015 in the text but the reference list gives 2016; please align the citation year with the reference list.
[Figure 3] The figure caption says "Example images and multi labels of Rescuenet dataset," but the figure does not include a scale bar or a description of the color coding for the multi-label masks, which makes it difficult for the reader to interpret the examples.
[Data section, first paragraph] The paper does not state how many classes are used in the multi-label classification task; the RescueNet dataset includes multiple instance types and damage levels, but the exact class set used for training and evaluation is not enumerated. This should be specified, for example in a table listing all classes and their positive-image counts.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the 90.23% mAP result is an external benchmark comparison on held-out RescueNet data, and the CSRA module is cited from independent prior work.

full rationale

The paper's central claim is an empirical benchmark result: ResNet-152 plus the CSRA attention module reaches 90.23% mAP on the public RescueNet dataset, compared with baseline models. The evaluation uses an 80/20 train/test split, so the reported mAP is computed on held-out images and is not equivalent by construction to any fitted parameter or training label. The CSRA module is adopted from Zhu and Wu (2021), an external prior work with no overlap with the present authors, and the feature extraction backbones are standard ResNet, VGG, and EfficientNet architectures. No equation in the paper defines the prediction in terms of the ground-truth labels, and no fitted parameter is renamed as a prediction. The paper does not invoke a uniqueness theorem or rely on a self-citation chain to justify its central claim. The main limitation noted in the data description is that the conversion from RescueNet's pixel-level semantic segmentation labels to image-level multi-label targets is not documented; this is a reproducibility and validity concern about the input labels, not a circularity in the derivation. Therefore, no specific circular step can be identified under the standard criteria, and the appropriate finding is no significant circularity.

Assumptions & free parameters 1 free parameters · 3 assumptions · 0 invented entities

The central claim is an empirical benchmark that depends on the correctness of the RescueNet labels and on the unstated derivation of multi-label ground truth from pixel-level annotations. The network weights themselves are learned, not free parameters; the only hand-chosen hyperparameter with a concrete value is the decision threshold. No new entities are postulated.

free parameters (1)

Decision threshold = 0.5
A fixed 0.5 threshold is applied to sigmoid outputs to derive binary labels. This is a hand-chosen hyperparameter, but it does not affect the reported mAP, which is threshold-independent.

assumptions (3)

domain assumption RescueNet labels accurately reflect ground-truth damage.
The evaluation treats the provided labels as correct. Any labeling errors or ambiguity in the dataset would propagate into the reported mAP. The paper does not analyze label quality.
ad hoc to paper Image-level multi-label ground truth can be reliably derived from pixel-level semantic segmentation labels.
RescueNet provides pixel-level segmentation, but the paper does not explain how these were aggregated to image-level labels. This assumption is unstated yet necessary for training and evaluation.
standard math The CSRA module from Zhu and Wu (2021) operates as described in the original paper.
The paper relies on the correctness of the cited CSRA formulation without re-deriving it.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Multi-Label Classification Framework for Hurricane Damage Assessment." pith.science (2026). https://pith.science/paper/NVRBRXGA

@misc{pith2026250702265,
  author       = {Pith},
  title        = {Pith review of: Multi-Label Classification Framework for Hurricane Damage Assessment},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/NVRBRXGA}},
  note         = {Machine review of arXiv:2507.02265}
}

read the original abstract

Hurricanes cause widespread destruction, resulting in diverse damage types and severities that require timely and accurate assessment for effective disaster response. While traditional single-label classification methods fall short of capturing the complexity of post-hurricane damage, this study introduces a novel multi-label classification framework for assessing damage using aerial imagery. The proposed approach integrates a feature extraction module based on ResNet and a class-specific attention mechanism to identify multiple damage types within a single image. Using the Rescuenet dataset from Hurricane Michael, the proposed method achieves a mean average precision of 90.23%, outperforming existing baseline methods. This framework enhances post-hurricane damage assessment, enabling more targeted and efficient disaster response and contributing to future strategies for disaster mitigation and resilience. This paper has been accepted at the ASCE International Conference on Computing in Civil Engineering (i3CE 2025), and the camera-ready version will appear in the official conference proceedings.

Figures

Figures reproduced from arXiv: 2507.02265 by the authors.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

FloodVision: Urban Flood Depth Estimation Using Foundation Vision-Language Models and Domain Knowledge Graph
cs.CV 2025-09 conditional novelty 5.0 of 10

FloodVision uses GPT-4o plus a knowledge graph of object heights to estimate urban flood depth from RGB images, achieving 8.17 cm MAE on 110 crowdsourced images.

Reference graph

Works this paper leans on

17 extracted references · 14 canonical work pages · cited by 1 Pith paper

[6]

Relation Network for Multilabel Aerial Image Classification

“Relation Network for Multilabel Aerial Image Classification.” IEEE Trans. Geosci. Remote Sens., 58(7): 4558–4572. https://doi.org/10.1109/TGRS.2019.2963364. Jing, R., S. Heft-Neal, D. R. Chavas, M. Griswold, Z. Wang, A. Clark-Ginsberg, D. Guha-Sapir, E. Bendavid, and Z. Wagner

work page arXiv 2019
[9]

Post-disaster damage classification based on deep multi-view image fusion

“Post-disaster damage classification based on deep multi-view image fusion.” Computer-Aided Civil and Infrastructure Engineering, 38(4): 528–544. https://doi.org/10.1111/mice.12890. Liu, C., S. M. E. Sepasgozar, Q. Zhang, and L. Ge

work page doi:10.1111/mice.12890
[11]

Mapping hurricane damage: A comparative analysis of satellite monitoring methods

“Mapping hurricane damage: A comparative analysis of satellite monitoring methods.” International Journal of Applied Earth Observation and Geoinformation, 91: 102134. https://doi.org/10.1016/j.jag.2020.102134. Paszke, A., S. Gross, F. Massa, A. Lerer, J. Bradbury, G. Chanan, T. Killeen, Z. Lin, N. Gimelshein, L. Antiga, A. Desmaison, A. Köpf, E. Yang, Z. ...

work page arXiv 2020
[13]

Global Context-Based Multilevel Feature Fusion Networks for Multilabel Remote Sensing Image Scene Classification

“Global Context-Based Multilevel Feature Fusion Networks for Multilabel Remote Sensing Image Scene Classification.” IEEE J. Sel. Top. Appl. Earth Obs. Remote Sens., 14: 11179–11196. https://doi.org/10.1109/JSTARS.2021.3122464. 9 Wang, Y., A. W. Z. Chew, and L. Zhang

work page arXiv 2021
[14]

Building damage detection from satellite images after natural disasters on extremely imbalanced datasets

“Building damage detection from satellite images after natural disasters on extremely imbalanced datasets.” Automation in Construction, 140: 104328. https://doi.org/10.1016/j.autcon.2022.104328. Weinkle, J., C. Landsea, D. Collins, R. Musulin, R. P. Crompton, P. J. Klotzbach, and R. Pielke

arXiv 2022
[16]

Mortality caused by tropical cyclones in the United States

“Mortality caused by tropical cyclones in the United States.” Nature, 635(8037): 121–128. https://doi.org/10.1038/s41586-024-07945-5. Zhou, W., P. Dou, T. Su, H. Hu, and Z. Zheng

work page doi:10.1038/s41586-024-07945-5
[17]

Feature learning network with transformer for multi-label image classification

“Feature learning network with transformer for multi-label image classification.” Pattern Recognition, 136: 109203. https://doi.org/10.1016/j.patcog.2022.109203. Zhu, K., and Wu, J. (2021). “Residual Attention: A Simple but Effective Method for Multi-Label Recognition.” Proc., IEEE/CVF International Conference on Computer Vision (ICCV), 184–193

work page arXiv 2021
[913]

Wang, X., L

https://doi.org/10.1038/s41597-023-02799-4. Wang, X., L. Duan, and C. Ning

doi:10.1038/s41597-023-02799-4

Show all 17 references

[2016]

Deep Residual Learning for Image Recognition

“Deep Residual Learning for Image Recognition.” Proc., IEEE Conf. on Computer Vision and Pattern Recognition (CVPR), 770–778. https://doi.org/10.1109/CVPR.2016.90. Hua, Y., L. Mou, and X. X. Zhu

2016 doi
[2017]

Structural analysis of the relationships between implementing tasks and ex-post effectiveness for disaster mitigation practice in Taipei City

“Structural analysis of the relationships between implementing tasks and ex-post effectiveness for disaster mitigation practice in Taipei City.” Sustainable Cities and Society, 34: 1–11. https://doi.org/10.1016/j.scs.2017.06.002. 8 He, K., X. Zhang, S. Ren, and J. Sun

2017 doi
[2018]

Normalized hurricane damage in the continental United States 1900–2017

“Normalized hurricane damage in the continental United States 1900–2017.” Nat. Sustain., 1(12): 808–813. https://doi.org/10.1038/s41893-018-0165-2. Young, R., and S. Hsiang

1900 doi
[2020]

Building damage annotation on post-hurricane satellite imagery based on convolutional neural networks

“Building damage annotation on post-hurricane satellite imagery based on convolutional neural networks.” Nat. Hazards, 103(3): 3357–3376. https://doi.org/10.1007/s11069-020-04133-2. Chou, J.-S., S.-C. Hsu, C.-W. Lin, C.-J. Lee, and J.-L. Wei

doi
[2021]

Hurricane Damage Detection using Machine Learning and Deep Learning Techniques: A Review

“Hurricane Damage Detection using Machine Learning and Deep Learning Techniques: A Review.” IOP Conf. Ser.: Mater. Sci. Eng., 1022(1): 012035. https://doi.org/10.1088/1757-899X/1022/1/012035. Khajwal, A. B., C.-S. Cheng, and A. Noshadravan

doi
[2022]

A novel attention-based deep learning method for post-disaster building damage classification

“A novel attention-based deep learning method for post-disaster building damage classification.” Expert Systems with Applications, 202: 117268. https://doi.org/10.1016/j.eswa.2022.117268. Lozano, J., & Tien, I. (2023). Data collection tools for post-disaster damage assessment ...

2023
[2023]

Increased U.S. coastal hurricane risk under climate change

“Increased U.S. coastal hurricane risk under climate change.” Science Advances, 9(14): eadf0259. https://doi.org/10.1126/sciadv.adf0259. Cao, Q. D., and Y. Choe

doi
[2024]

Global population profile of tropical cyclone exposure from 2002 to 2019

“Global population profile of tropical cyclone exposure from 2002 to 2019.” Nature, 626(7999): 549–554. https://doi.org/10.1038/s41586-023-06963-z. Kaur, S., S. Gupta, and S. Singh

2002 doi
[2344]

Balaguru, K., W

https://doi.org/10.3390/buildings14082344. Balaguru, K., W. Xu, C.-C. Chang, L. R. Leung, D. R. Judi, S. M. Hagos, M. F. Wehner, J. P. Kossin, and M. Ting

doi

Pith tools

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

[1] [6]

Relation Network for Multilabel Aerial Image Classification

“Relation Network for Multilabel Aerial Image Classification.” IEEE Trans. Geosci. Remote Sens., 58(7): 4558–4572. https://doi.org/10.1109/TGRS.2019.2963364. Jing, R., S. Heft-Neal, D. R. Chavas, M. Griswold, Z. Wang, A. Clark-Ginsberg, D. Guha-Sapir, E. Bendavid, and Z. Wagner

work page arXiv 2019

[2] [9]

Post-disaster damage classification based on deep multi-view image fusion

“Post-disaster damage classification based on deep multi-view image fusion.” Computer-Aided Civil and Infrastructure Engineering, 38(4): 528–544. https://doi.org/10.1111/mice.12890. Liu, C., S. M. E. Sepasgozar, Q. Zhang, and L. Ge

work page doi:10.1111/mice.12890

[3] [11]

Mapping hurricane damage: A comparative analysis of satellite monitoring methods

“Mapping hurricane damage: A comparative analysis of satellite monitoring methods.” International Journal of Applied Earth Observation and Geoinformation, 91: 102134. https://doi.org/10.1016/j.jag.2020.102134. Paszke, A., S. Gross, F. Massa, A. Lerer, J. Bradbury, G. Chanan, T. Killeen, Z. Lin, N. Gimelshein, L. Antiga, A. Desmaison, A. Köpf, E. Yang, Z. ...

work page arXiv 2020

[4] [13]

Global Context-Based Multilevel Feature Fusion Networks for Multilabel Remote Sensing Image Scene Classification

“Global Context-Based Multilevel Feature Fusion Networks for Multilabel Remote Sensing Image Scene Classification.” IEEE J. Sel. Top. Appl. Earth Obs. Remote Sens., 14: 11179–11196. https://doi.org/10.1109/JSTARS.2021.3122464. 9 Wang, Y., A. W. Z. Chew, and L. Zhang

work page arXiv 2021

[5] [14]

Building damage detection from satellite images after natural disasters on extremely imbalanced datasets

“Building damage detection from satellite images after natural disasters on extremely imbalanced datasets.” Automation in Construction, 140: 104328. https://doi.org/10.1016/j.autcon.2022.104328. Weinkle, J., C. Landsea, D. Collins, R. Musulin, R. P. Crompton, P. J. Klotzbach, and R. Pielke

arXiv 2022

[6] [16]

Mortality caused by tropical cyclones in the United States

“Mortality caused by tropical cyclones in the United States.” Nature, 635(8037): 121–128. https://doi.org/10.1038/s41586-024-07945-5. Zhou, W., P. Dou, T. Su, H. Hu, and Z. Zheng

work page doi:10.1038/s41586-024-07945-5

[7] [17]

Feature learning network with transformer for multi-label image classification

“Feature learning network with transformer for multi-label image classification.” Pattern Recognition, 136: 109203. https://doi.org/10.1016/j.patcog.2022.109203. Zhu, K., and Wu, J. (2021). “Residual Attention: A Simple but Effective Method for Multi-Label Recognition.” Proc., IEEE/CVF International Conference on Computer Vision (ICCV), 184–193

work page arXiv 2021

[8] [913]

Wang, X., L

https://doi.org/10.1038/s41597-023-02799-4. Wang, X., L. Duan, and C. Ning

doi:10.1038/s41597-023-02799-4