Pith. sign in

REVIEW 4 major objections 4 minor 37 references

Damage Assessment after Natural Disasters with UAVs: Semantic Feature Extraction using Deep Learning

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

Pith's one-line read A learnable semantic filter lets UAVs send over 85% less data without hurting downstream accuracy.

desk verdict Plausible new idea—jointly trained binary masking over semantic masks—but the paper's own RescueNet numbers contradict its central accuracy claim, and the bandwidth win is mostly from the segmentation mask, not the new mask. read the letter →

arxiv 2412.10756 v1 pith:KIL7YTDT submitted 2024-12-14 cs.CV cs.LG

classification cs.CVcs.LG
keywords disasterassessmentunmannedaerialvehiclessemanticsegmentationdeeplearningvisualquestionansweringbandwidth-efficientcommunicationbinarymaskingfeatureextraction
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

After a natural disaster, UAVs need to send imagery to ground stations, but bandwidth is scarce and unreliable. This paper proposes a learnable 'semantic extractor' that runs onboard the UAV: it turns the image into a semantic segmentation map, then predicts a binary mask that keeps only regions relevant to the specific downstream task, and transmits their element-wise product $y = M \odot B$. The authors claim that on the FloodNet and RescueNet benchmarks this reduces transmitted data by more than 85% while keeping the downstream tasks' accuracy at an acceptable level: VQA error is essentially unchanged, and damage-classification error increases by an amount the paper accepts as a trade-off for the bandwidth gain. If true, disaster-response teams could receive situational awareness faster without sacrificing the decisions that matter.

What carries the argument

The central object is the binary mask predictor, a fully convolutional network with three transposed convolutional layers, a $1\times 1$ convolution, and a Gumbel-Softmax activation that maps the PSPNet semantic mask $M$ to a binary mask $B$. The transmitted signal is the element-wise product $y = M \odot B$. The mechanism is trained jointly with the downstream model: a sparsity loss $L_{\text{sparsity}} = \frac{1}{N}\sum_{i=1}^{N}|y_i - p|$ (with $p$ a zero matrix) pushes the mask toward fewer regions, while the downstream task's categorical cross-entropy loss keeps the filtered representation informative. Gumbel-Softmax makes the binary selection differentiable, so gradients flow from the downstream task back through the mask.

What would settle it

A direct test is to take the RescueNet classification task and evaluate the masked semantic map against the original image at equal transmission cost by also compressing the original image to the same byte budget; if the compressed original matches or beats the masked map's 41.33% error, the binary mask is not the source of the bandwidth savings.

Watch

Extended reading notes

Core claim

The paper's central claim is that a task-conditioned binary mask can filter a semantic segmentation mask so that the transmitted representation retains the information a given downstream model needs. The binary mask is produced by a small fully convolutional network with a Gumbel-Softmax activation, trained jointly with the downstream model using a weighted sum of a sparsity loss and the downstream task's categorical cross-entropy loss. The transmitted item is the element-wise product $y = M \odot B$ of the PSPNet semantic mask $M$ and the binary mask $B$. On FloodNet visual question answering, the masked representation achieves an overall error of 31.00% versus 31.11% for the original image; on RescueNet damage classification, error rises from 30.00% for the original image to 41.33% for the masked map. The authors interpret these results as maintaining downstream performance while cutting transmitted data size by roughly 86% on FloodNet and 92% on RescueNet, with corresponding reductions in transmission latency.

Load-bearing premise

The approach assumes the semantic segmentation mask already contains all decision-critical information, so filtering it cannot remove a detail the downstream task needs.

Editorial extensions

If this is right

  • If the claim holds, disaster-response UAVs can transmit masked semantic maps instead of full images over narrow links, cutting per-image latency by roughly 86–92% in the free-space path-loss scenarios modeled here.
  • The binary mask can be retrained for any downstream model, so the same segmentation module could serve VQA, damage-level classification, or future analytics without changing the transmission format.
  • Because the mask predictor adds only about 0.044 million parameters, the filtering step is light enough to run onboard a UAV alongside PSPNet, making the bandwidth savings available in real time.
  • For the Yes/No question type on FloodNet, the masked input actually reduces error relative to the original image (24.44% vs 38.89%), suggesting that removing irrelevant detail can help simple decisions.
  • The 86–92% reduction in transmitted bytes translates directly into lower transmission latency under the paper's link model, meaning responders can receive updates sooner.

Reading between the lines

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

  • Beyond the paper, the same masking idea could be applied to learned feature maps or embeddings, selecting spatial regions rather than semantic classes; the paper does not test this variant.
  • The RescueNet result—error rising from 30.00% with the original image to 41.33% with the masked map—suggests the load-bearing assumption that the semantic mask preserves decision-critical detail is weaker for fine-grained damage grading than for counting and condition questions; evaluating the mask on tasks that need sub-class texture information would test this.
  • A testable extension is to compare the masked semantic map against transmitting the full semantic map at the same bit budget; if the unfiltered map yields similar accuracy with similar data volume, the binary mask's value would be prioritisation rather than compression.
  • The latency model assumes a line-of-sight free-space channel; under occlusion or multi-path conditions the relative benefit of smaller payloads would shrink, so practical gains in cluttered disaster environments may differ from the reported numbers.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 4 minor

Summary. The paper proposes a bandwidth-reduction pipeline for UAV-based disaster damage assessment. An onboard PSPNet semantic segmentation model produces a class mask, and a small FCN with a Gumbel-Softmax (or Sigmoid, per Fig. 3) output predicts a binary importance mask; the transmitted signal is the elementwise product of the semantic mask and the binary mask, y = M ⊙ B (Eq. 10). The binary-mask predictor is trained jointly with a downstream model using a weighted sum of an L1 sparsity loss and the task loss (Eq. 14). The method is evaluated on FloodNet for visual question answering and RescueNet for building damage classification, with comparisons among the original image, ground-truth mask, predicted semantic mask, and masked semantic mask. The paper claims a >85% reduction in transmitted data while maintaining downstream-task accuracy.

Significance. If the central claim were established, the paper would address a practical bottleneck in UAV-enabled disaster response: limited bandwidth between the UAV and ground station. The work has several strengths: it uses two public benchmark datasets, the proposed mask predictor is very small (0.044M parameters, Table VI), the latency model is clearly described, and the paper explicitly attempts to isolate the effect of the binary mask by comparing against the predicted semantic mask alone. However, the quantitative evidence does not support the headline claim of maintained accuracy: on RescueNet the proposed masked segmentation mask increases classification error to 41.33%, versus 30.00% for the original image and 36.00% for the semantic mask alone (Table II). The claimed >85% data reduction is also mostly attributable to replacing the image with a semantic segmentation mask rather than to the novel learned binary mask. The significance of the contribution is therefore currently unestablished, although the overall idea is of interest to the semantic-communication and UAV-perception communities.

major comments (4)
  1. [§IV-C, Table II] The central claim in the abstract and contribution 3, that the method 'maintains high accuracy' and 'maintaining the performance in downstream tasks,' is contradicted by the RescueNet results. Table II reports classification error of 30.00% for the original image, 36.00% for the predicted semantic mask, and 41.33% for the proposed masked semantic mask. The learned binary mask therefore increases error by 5.33 percentage points relative to the semantic mask alone and by 11.33 points relative to the original image. The text states that this increase is 'not significant,' but the paper reports no error bars, confidence intervals, or significance tests. For a disaster-damage triage task on a roughly 30% error baseline, a 5.3-point degradation is material, and the claim of maintained performance fails for one of the two demonstrated tasks.
  2. [§IV-C, Tables III and IV] The headline '>85% reduction of the transmitted data' is dominated by the semantic segmentation baseline, not by the novel binary mask. On FloodNet, going from the full image (14.441 kB) to the predicted semantic mask (2.119 kB) already achieves an 85.3% reduction; the incremental reduction from the binary mask is only 8.2% (2.119 kB to 1.945 kB). On RescueNet, the image-to-semantic-mask step gives an 84.5% reduction (177.78 kB to 27.559 kB), and the incremental mask reduction is 50.2% (27.559 kB to 13.728 kB). Since transmitting a semantic segmentation mask is an existing approach cited in the related work, the claims in the abstract and contribution 3 overstate the contribution of the proposed learnable extractor.
  3. [§IV-B and Table V] The Jaccard index and MSE fidelity evidence in Table V is computed using feature activations of the same downstream model that is jointly trained with the binary mask through the loss in Eq. (14). High Jaccard and low MSE on the test set therefore only show that the mask does not dramatically alter the internal representation of this particular trained model; they do not provide an independent measure of whether decision-critical information is preserved. This is especially problematic because Table II shows a clear loss of classifiable information for RescueNet despite the near-unity Jaccard index of 0.993. The fidelity metric does not rescue the central accuracy claim.
  4. [§III-B and Fig. 3] There is a reproducibility-relevant inconsistency in the description of the binary mask predictor. Section III-B and Eq. (9) state that a Gumbel-Softmax activation φ is applied to the single-channel output, while Fig. 3 shows a Sigmoid followed by resizing. Gumbel-Softmax is normally defined over a categorical distribution, not a single logit, so the exact discrete relaxation used needs to be specified. As written, the architecture cannot be reimplemented unambiguously from the text and figure.
minor comments (4)
  1. [§III-D, Eq. (12)] The variable name Lsparcity contains a typo and should read Lsparsity; also p is described only as 'a matrix of zeros,' but the size of p should be defined explicitly.
  2. [§IV, first paragraph] The phrase 'feasible to be deployed onbard of a UA V' contains a typo: 'onbard' should be 'onboard.'
  3. [Table V caption] The caption says 'The matrices are calculated,' but the paper reports scalar Jaccard and MSE values; 'The metrics are calculated' is the intended wording.
  4. [§IV-B, Eqs. (15)-(16)] The data sizes in Tables III and IV are surprisingly small for 3000×4000 RGB images unless a specific compression format and bit depth are assumed; the paper should state the encoding (e.g., JPEG quality or PNG bit depth) used for the reported average data sizes.

Circularity Check

1 steps flagged · score 3.0 of 10

The central accuracy claim is empirical and not circular, but the supporting semantic-fidelity evidence in Table V is self-referential because the binary mask and downstream model were jointly trained, so high Jaccard and low MSE reflect the training objective rather than an independent check.

  1. fitted input called prediction [Section III-D (Eq. 14) and Section IV-C (Table V and following paragraph)]
    "The loss for training the binary mask predicting model Loss can be calculated by Loss = wsLsparcity + wcLcategorical , ... the training of the binary mask predictor occurs jointly with the downstream model. ... This clearly indicates that the binary mask has preserved the critical information that is required for decision-making and discarded the uninformative content."

    Eq. (14) trains the binary mask B and the downstream model jointly by minimizing ws*Lsparcity + wc*Lcategorical, so B is explicitly optimized to let the jointly trained model retain accuracy after masking. Table V then measures feature overlap between the predicted mask and the masked mask using that same jointly trained model, and the near-1 Jaccard / near-0 MSE is presented as evidence that the binary mask has preserved critical information. This is not an independent test: high feature agreement is the expected trace of co-optimizing the mask and the model, so the claimed semantic-fidelity result reduces to the fitted loss rather than to an external source of information.

full rationale

The paper does not contain a formal derivation chain whose output equals its input; its headline result is an empirical comparison of error rates across input representations. Tables I and II test the same downstream architectures on original images, predicted semantic masks, and the proposed masked mask, so the 'maintains performance' claim is anchored to an external baseline rather than forced by construction. The one genuinely self-referential piece of evidence is Table V: because B and the downstream model are trained jointly under Eq. (14), measuring feature similarity with that same model is a diagnostic of the training objective, not an independent confirmation that y retains decision-critical content. That is a minor circularity in the supporting evidence. Separately, the RescueNet row of Table II shows error rising from 30.00% (original image) to 36.00% (predicted semantic mask) to 41.33% (masked mask), which the paper calls 'not significant' without reporting error bars or significance tests; this is an empirical weakness and does not by itself make the derivation circular. The loss weights ws and wc are also said to be 'experimentally evaluated', which creates a tuning risk but is not a circular reduction. Overall, the central claim is not circular, but the auxiliary fidelity argument is self-referential, so the score is low rather than zero.

Assumptions & free parameters 2 free parameters · 3 assumptions · 0 invented entities

No new physical entities are postulated. The semantic extractor is a learned neural component, not an invented entity. The main free parameters are the loss weights and Gumbel temperature, both chosen by hand without reported values.

free parameters (2)
  • Loss weights w_s and w_c = not reported
    Equation (14) combines sparsity and downstream losses with weights 'experimentally evaluated' (Section III-D), but values are not given; they control the accuracy-data tradeoff.
  • Gumbel-Softmax temperature = not reported
    Section III-B applies Gumbel-Softmax for differentiable binary selection; the temperature is not specified, affecting how hard or soft the mask is.
assumptions (3)
  • domain assumption The PSPNet semantic segmentation mask M contains sufficient information for the downstream VQA and damage classification tasks.
    The transmitted representation is M ⊙ B (Eq. 10), so any information absent from M is unavailable to ground-station models.
  • domain assumption The binary mask predictor can be trained end-to-end with the downstream model and sparsity loss to select task-relevant regions without degrading task performance.
    Central mechanism in Section III-D; the validity is exactly what the experiments test, but is assumed by the framework.
  • domain assumption The free-space path loss model for the UAV-to-ground link (Eqs. 15-16) approximates real disaster communication channels.
    Used to compute latency; not central to the accuracy claim but supports the significance argument.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Damage Assessment after Natural Disasters with UAVs: Semantic Feature Extraction using Deep Learning." pith.science (2026). https://pith.science/paper/KIL7YTDT

@misc{pith2026241210756,
  author       = {Pith},
  title        = {Pith review of: Damage Assessment after Natural Disasters with UAVs: Semantic Feature Extraction using Deep Learning},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/KIL7YTDT}},
  note         = {Machine review of arXiv:2412.10756}
}
read the original abstract

Unmanned aerial vehicle-assisted disaster recovery missions have been promoted recently due to their reliability and flexibility. Machine learning algorithms running onboard significantly enhance the utility of UAVs by enabling real-time data processing and efficient decision-making, despite being in a resource-constrained environment. However, the limited bandwidth and intermittent connectivity make transmitting the outputs to ground stations challenging. This paper proposes a novel semantic extractor that can be adopted into any machine learning downstream task for identifying the critical data required for decision-making. The semantic extractor can be executed onboard which results in a reduction of data that needs to be transmitted to ground stations. We test the proposed architecture together with the semantic extractor on two publicly available datasets, FloodNet and RescueNet, for two downstream tasks: visual question answering and disaster damage level classification. Our experimental results demonstrate the proposed method maintains high accuracy across different downstream tasks while significantly reducing the volume of transmitted data, highlighting the effectiveness of our semantic extractor in capturing task-specific salient information.

Figures

Figures reproduced from arXiv: 2412.10756 by the authors.

Figure 1
Figure 1. An overview of the architecture of the proposed framework. First, the emergency area is captured using a UAV and it is converted into a semantic [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 3
Figure 3. The architecture of the data masking model. Initially, the segmented [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figure 5
Figure 5. The architecture of the proposed classifier model. First, the image [PITH_FULL_IMAGE:figures/full_fig_p005_5.png] view at source ↗
Figures from the paper (3 more)
Figure 6
Figure 6. Figure 6: Samples visualisations of the original image, ground segmentation truth mask, predicted segmentation mask, and masked segmentation mask for the [PITH_FULL_IMAGE:figures/full_fig_p008_6.png]
Figure 7
Figure 7. Figure 7: Samples visualisations of the original image, ground segmentation truth mask, predicted segmentation mask, and masked segmentation mask for the [PITH_FULL_IMAGE:figures/full_fig_p009_7.png]
Figure 8
Figure 8. Figure 8: Class wise mIoU on 0 to 1 scale between predicted masks using PSPNet and ground truth masks for datasets (a) Floodnet. (b) Rescuenet. [PITH_FULL_IMAGE:figures/full_fig_p010_8.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

37 extracted references · 32 canonical work pages

  1. [1]

    WMO Atlas of mortal- ity and economic losses from weather, climate, and water extremes (1970–2019),

    World Meteorological Organization (WMO), “WMO Atlas of mortal- ity and economic losses from weather, climate, and water extremes (1970–2019),” 2021

  2. [2]

    GAR special report 2023: Mapping resilience for the sustainable development goals,

    United Nations Office for Disaster Risk Reduction (UNDRR), “GAR special report 2023: Mapping resilience for the sustainable development goals,” 2023

  3. [3]

    Solutions for sus- tainable and resilient communication infrastructure in disaster relief and management scenarios,

    B. Karaman, I. Basturk, S. Taskin, E. Zeydan, F. Kara, E. A. Beyazit, M. Camelo, E. Bj ¨ornson, and H. Yanikomeroglu, “Solutions for sus- tainable and resilient communication infrastructure in disaster relief and management scenarios,” arXiv preprint arXiv:2410.13977 , 2024

  4. [4]

    Onboard radar processor development for rapid response to natural hazards,

    Y . Lou, D. Clark, P. Marks, R. J. Muellerschoen, and C. C. Wang, “Onboard radar processor development for rapid response to natural hazards,” IEEE J. Sel. Topics Appl. Earth Observ. Remote Sens. , vol. 9, no. 6, pp. 2770–2776, 2016

  5. [5]

    UA V computing-assisted search and rescue mission framework for disaster and harsh environment mitigation,

    S. H. Alsamhi, A. V . Shvetsov, S. Kumar, S. V . Shvetsova, M. A. Alhartomi, A. Hawbani, N. S. Rajput, S. Srivastava, A. Saif, and V . O. Nyangaresi, “UA V computing-assisted search and rescue mission framework for disaster and harsh environment mitigation,” Drones, vol. 6, no. 7, p. 154, 2022. xi

  6. [6]

    A uav-assisted edge framework for real-time disaster management,

    H. Ijaz, R. Ahmad, R. Ahmed, W. Ahmed, Y . Kai, and W. Jun, “A uav-assisted edge framework for real-time disaster management,” IEEE Transactions on Geoscience and Remote Sensing , vol. 61, pp. 1–13, 2023

  7. [7]

    Ugen: Uav and gan-aided ensemble network for post-disaster survivor detection through oran,

    G. Raja, A. Manoharan, and H. Siljak, “Ugen: Uav and gan-aided ensemble network for post-disaster survivor detection through oran,” IEEE Transactions on Vehicular Technology , vol. 73, no. 7, pp. 9296– 9305, 2024

  8. [8]

    Semantic communications: Overview, open issues, and future research directions,

    X. Luo, H.-H. Chen, and Q. Guo, “Semantic communications: Overview, open issues, and future research directions,” IEEE Wirel. Commun. , vol. 29, no. 1, pp. 210–219, 2022

Show all 37 references
  1. [9]

    Wireless end-to-end image transmission system using semantic communications,

    M. U. Lokumarambage, V . S. S. Gowrisetty, H. Rezaei, T. Sivalingam, N. Rajatheva, and A. Fernando, “Wireless end-to-end image transmission system using semantic communications,” IEEE Access , vol. 11, pp. 37 149–37 163, 2023

  2. [10]

    Semantic segmentation for high spatial resolution remote sensing images based on convolution neural network and pyramid pooling module,

    B. Yu, L. Yang, and F. Chen, “Semantic segmentation for high spatial resolution remote sensing images based on convolution neural network and pyramid pooling module,” IEEE J. Sel. Topics Appl. Earth Observ. Remote Sens., vol. 11, no. 9, pp. 3252–3261, 2018

  3. [11]

    A sim2real deep learning ap- proach for the transformation of images from multiple vehicle-mounted cameras to a semantically segmented image in bird’s eye view,

    L. Reiher, B. Lampe, and L. Eckstein, “A sim2real deep learning ap- proach for the transformation of images from multiple vehicle-mounted cameras to a semantically segmented image in bird’s eye view,” in Proc. IEEE 23rd Intl. Conf. Intelligent Transportation Systems (ITSC) , 2...

  4. [12]

    Lightweight disaster semantic segmentation for uav on-device intelligence,

    H. Lee, G. Kim, S. Ha, and H. Kim, “Lightweight disaster semantic segmentation for uav on-device intelligence,” in IGARSS 2024 - 2024 IEEE International Geoscience and Remote Sensing Symposium , 2024, pp. 8821–8825

  5. [13]

    Floodnet: A high resolution aerial imagery dataset for post flood scene understanding,

    M. Rahnemoonfar, T. Chowdhury, A. Sarkar, D. Varshney, M. Yari, and R. R. Murphy, “Floodnet: A high resolution aerial imagery dataset for post flood scene understanding,” IEEE Access, vol. 9, pp. 89 644–89 654, 2021

  6. [14]

    Rescuenet: A high resolution UA V semantic segmentation dataset for natural disaster damage assessment,

    M. Rahnemoonfar, T. Chowdhury, and R. Murphy, “Rescuenet: A high resolution UA V semantic segmentation dataset for natural disaster damage assessment,” Scientific Data, vol. 10, no. 1, p. 913, 2023

  7. [15]

    Lsar: Multi-uav col- laboration for search and rescue missions,

    E. T. Alotaibi, S. S. Alqefari, and A. Koubaa, “Lsar: Multi-uav col- laboration for search and rescue missions,” IEEE Access , vol. 7, pp. 55 817–55 832, 2019

  8. [16]

    Uav aerial imaging applications for post-disaster assessment, environmental management and infrastructure development,

    C. A. F. Ezequiel, M. Cua, N. C. Libatique, G. L. Tangonan, R. Alampay, R. T. Labuguen, C. M. Favila, J. L. E. Honrado, V . Ca ˜nos, C. De- vaney, A. B. Loreto, J. Bacusmo, and B. Palma, “Uav aerial imaging applications for post-disaster assessment, environmental management an...

  9. [18]

    Uav-based real-time survivor detection system in post-disaster search and rescue operations,

    J. Dong, K. Ota, and M. Dong, “Uav-based real-time survivor detection system in post-disaster search and rescue operations,” IEEE Journal on Miniaturization for Air and Space Systems , vol. 2, no. 4, pp. 209–219, 2021

  10. [19]

    Drones4good: Supporting disaster relief through remote sens- ing and ai,

    N. Merkle, R. Bahmanyar, C. Henry, S. M. Azimi, X. Yuan, S. Schopferer, V . Gstaiger, S. Auer, A. Schneibel, M. Wieland, and T. Kraft, “Drones4good: Supporting disaster relief through remote sens- ing and ai,” in 2023 IEEE/CVF International Conference on Computer Vision Worksh...

  11. [20]

    Deepdamagenet: A two-step deep-learning model for multi-disaster building damage segmentation and classification using satellite imagery,

    I. Alisjahbana, J. Li, Ben, Strong, and Y . Zhang, “Deepdamagenet: A two-step deep-learning model for multi-disaster building damage segmentation and classification using satellite imagery,” 2024

  12. [21]

    Transferring cnn with adaptive learning for remote sensing scene classification,

    W. Wang, Y . Chen, and P. Ghamisi, “Transferring cnn with adaptive learning for remote sensing scene classification,” IEEE Transactions on Geoscience and Remote Sensing , vol. 60, pp. 1–18, 2022

  13. [22]

    Deep-learning-based aerial image classification for emergency response applications using unmanned aerial vehicles,

    C. Kyrkou and T. Theocharides, “Deep-learning-based aerial image classification for emergency response applications using unmanned aerial vehicles,” in 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) , 2019, pp. 517–525

  14. [23]

    TinyVQA: Compact multimodal deep neural network for visual question answering on resource-constrained devices,

    H.-A. Rashid, A. Sarkar, A. Gangopadhyay, M. Rahnemoonfar, and T. Mohsenin, “TinyVQA: Compact multimodal deep neural network for visual question answering on resource-constrained devices,” in Proc. tinyML Research Symposium – 2024 , 2024. [Online]. Available: https://arxiv.org...

  15. [24]

    Emergencynet: Efficient aerial image classification for drone-based emergency monitoring using atrous con- volutional feature fusion,

    C. Kyrkou and T. Theocharides, “Emergencynet: Efficient aerial image classification for drone-based emergency monitoring using atrous con- volutional feature fusion,” IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing , vol. 13, pp. 1687–1699, 2020

  16. [25]

    Visual AI for satellite imagery perspective: A visual question answering frame- work in the geospatial domain,

    S. Bhorge, M. Rane, N. Rane, M. Patil, P. Saraf, and J. Nilgar, “Visual AI for satellite imagery perspective: A visual question answering frame- work in the geospatial domain,” in Proc. 2023 IEEE 8th International Conference for Convergence in Technology (I2CT) , 2023, pp. 1–6

  17. [26]

    SAM-VQA: Supervised attention-based visual question answering model for post-disaster damage assessment on remote sensing imagery,

    A. Sarkar, T. Chowdhury, R. R. Murphy, A. Gangopadhyay, and M. Rah- nemoonfar, “SAM-VQA: Supervised attention-based visual question answering model for post-disaster damage assessment on remote sensing imagery,” IEEE Trans. Geosci. Remote Sens. , vol. 61, pp. 1–16, 2023

  18. [27]

    Unsupervised learning of image segmentation based on differentiable feature clustering,

    W. Kim, A. Kanezaki, and M. Tanaka, “Unsupervised learning of image segmentation based on differentiable feature clustering,” IEEE Trans. Image Process., vol. 29, pp. 8055–8068, 2020

  19. [28]

    Learning enriched features for fast image restoration and enhancement,

    S. W. Zamir, A. Arora, S. Khan, M. Hayat, F. S. Khan, M.-H. Yang, and L. Shao, “Learning enriched features for fast image restoration and enhancement,” IEEE Trans. Pattern Anal. Mach. Intell. , vol. 45, no. 2, pp. 1934–1948, 2023

  20. [29]

    Semantic communication systems for speech transmission,

    Z. Weng and Z. Qin, “Semantic communication systems for speech transmission,” IEEE J. Sel. Areas Commun. , vol. 39, no. 8, pp. 2434– 2444, 2021

  21. [30]

    Context-based semantic communication via dynamic programming,

    Y . Zhang, H. Zhao, J. Wei, J. Zhang, M. F. Flanagan, and J. Xiong, “Context-based semantic communication via dynamic programming,” IEEE Trans. Cogn. Commun. Netw., vol. 8, no. 3, pp. 1453–1467, 2022

  22. [31]

    UA V image high fidelity compression algorithm based on generative adversarial networks under complex disaster conditions,

    Q. Hu, C. Wu, Y . Wu, and N. Xiong, “UA V image high fidelity compression algorithm based on generative adversarial networks under complex disaster conditions,” IEEE Access , vol. 7, pp. 91 980–91 991, 2019

  23. [32]

    Blessemflood21: Advancing flood analysis with a high-resolution georeferenced dataset for humanitarian aid support,

    V . Polushko, A. Jenal, J. Bongartz, I. Weber, D. Hatic, R. R¨osch, T. M¨arz, M. Rauhut, and A. Weinmann, “Blessemflood21: Advancing flood analysis with a high-resolution georeferenced dataset for humanitarian aid support,” in Proc. 2024 IEEE International Geoscience and Remot...

  24. [33]

    Pyramid scene parsing network,

    H. Zhao, J. Shi, X. Qi, X. Wang, and J. Jia, “Pyramid scene parsing network,” in Proc. 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2017, pp. 6230–6239

  25. [34]

    Deep residual learning for image recognition,

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

  26. [35]

    An efficient modern baseline for FloodNet VQA,

    A. Kane and S. Khose, “An efficient modern baseline for FloodNet VQA,” in Proc. ICML 2022 New In Machine Learning (NewInML) Workshop, 2022. [Online]. Available: https://arxiv.org/pdf/2205.15025

  27. [36]

    RoBERTa: A robustly optimized BERT pretraining approach,

    Y . Liu, M. Ott, N. Goyal, J. Du, M. Joshi, D. Chen, O. Levy, M. Lewis, L. Zettlemoyer, and V . Stoyanov, “RoBERTa: A robustly optimized BERT pretraining approach,” 2019. [Online]. Available: https://arxiv.org/abs/1907.11692

  28. [37]

    Minimizing maximum latency of task offloading for multi-UA V-assisted maritime search and rescue,

    S. Qi, B. Lin, Y . Deng, X. Chen, and Y . Fang, “Minimizing maximum latency of task offloading for multi-UA V-assisted maritime search and rescue,” IEEE Trans. Veh. Technol. , vol. 73, no. 9, pp. 13 625–13 638, 2024

  29. [38]

    Efficient uavs deployment and resource allocation in uav-relay assisted public safety networks for video transmission,

    N. Khan, A. Ahmad, A. Wakeel, Z. Kaleem, B. Rashid, and W. Khalid, “Efficient uavs deployment and resource allocation in uav-relay assisted public safety networks for video transmission,” IEEE Access , vol. 12, pp. 4561–4574, 2024

Pith tools

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