Pith. sign in

REVIEW 4 major objections 7 minor 30 references

AeroLite-MDNet: Lightweight Multi-task Deviation Detection Network for UAV Landing

T0 review · 4 major / 7 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read AeroLite-MDNet claims a lightweight detection-plus-segmentation network can warn of UAV landing deviations in real time, with 98.6% warning accuracy and a 0.7-second average delay.

desk verdict The paper's new dataset and AWD metric are useful, but the reported 0.7s warning delay excludes inference latency, so the real-time claim is not supported as stated. read the letter →

arxiv 2506.21635 v1 pith:7SYJ3HCP submitted 2025-06-25 cs.RO cs.AIcs.CV

Haiping Yang , Huaxing Liu , Wei Wu , Zuohui Chen , Ning Wu This is my paper · ORCID
classification cs.ROcs.AIcs.CV
keywords UAVlandingdeviationwarningmulti-tasklearningobjectdetectionsemanticsegmentationlightweightnetworkedgedeploymentAverageDelay
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

Landing a UAV on a docking station often fails when GPS is weak or jammed, so pilots need a vision-based way to tell, in real time, that the aircraft is drifting off course. This paper proposes AeroLite-MDNet, a lightweight network that runs detection and segmentation together: detection finds the landing nest, QR code, or house, and segmentation recovers its shape, so the system can compare the target's position with the image center and decide whether to warn. To support this, the paper contributes UAVLandData, a 9,142-image landing dataset, and a new metric, Average Warning Delay (AWD), which measures how many seconds pass between the true start of a deviation and the system's warning. The reported result is 98.6% warning accuracy, a 0.7-second AWD, 7.12M parameters, 104.2 FPS on an RTX 3070, and 7.3 FPS on a Jetson TX2, which the authors present as evidence that real-time edge deployment is feasible.

What carries the argument

The carrying mechanism is a shared backbone-and-neck feature pyramid feeding two heads. The backbone uses CA Conv and SCA C3 attention modules to keep features compact while emphasizing relevant channels and spatial locations; the neck produces large, medium, and small feature maps. The detection head uses preset anchors with focal loss and CIoU loss, and the segmentation head fuses multi-scale features and combines Dice, BCEWithLogits, and focal losses, with the final mask cropped by the detected box to remove stray predictions. The decision rule (Algorithm 1) and the AWD metric complete the mechanism: AWD is the mean absolute difference between true deviation onset and warning time, with frames captured during model inference excluded so reported delay is not an artifact of processing speed.

What would settle it

Mount the camera with a fixed yaw offset so the landing pad is consistently, say, 40 pixels right of the image center at descent start, and run Algorithm 1 without recalibration; if the system still reports 98.6% warning accuracy and 0.7 s AWD on ground truth labeled by the true pad position, the center-alignment assumption is validated, and if not, the reported metrics are an artifact of that assumption.

Watch

Extended reading notes

Core claim

The central claim is that object detection and semantic segmentation can be combined in one small network to give a practical UAV landing-deviation warning. The detection branch supplies location and category for the landing target, and the segmentation branch supplies shape, which lets the system handle rotated targets that vertical bounding boxes cannot capture. Deviation is decided by Algorithm 1: if only a house is visible, the UAV deviates when the house is not at the image center; once the nest or QR code appears, the UAV deviates when the distance between the image center and target center divided by the nest's length exceeds a threshold. On the authors' data and baselines, the model reaches 98.6% warning accuracy and 0.7 seconds AWD while staying light enough (7.12M parameters, 104.2 FPS on GPU, 7.3 FPS on TX2) for near-real-time warning on embedded hardware.

Load-bearing premise

Everything hinges on the assumption that the center of the camera image is the true landing position and that the landing zone starts inside the center crop; if the camera is misaligned with the actual touchdown point, the labels and the reported warning delay measure the wrong quantity.

Editorial extensions

If this is right

  • A warning system of this size can run on embedded hardware (7.3 FPS on Jetson TX2), so landing monitoring need not depend on a ground-station GPU.
  • AWD gives the community a direct, interpretable measure of how fast a deviation warning arrives, not just whether the classification is correct.
  • The joint detection-segmentation design lets the system handle rotated and small landing targets better than either task alone, which matters in real descents.
  • The public dataset can serve as a benchmark for future vision-based landing-deviation research.

Reading between the lines

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

  • The system's reference point is the image center, so field deployments need camera alignment with the true touchdown point; adding a calibration or adaptive reference would make the numbers transferable to other aircraft.
  • AWD is a generalizable metric for any time-critical visual warning task where ground-truth onset times exist, not only UAV landings.
  • The warning output is a monitor, not a controller; wiring this signal into the flight control loop is the natural next step, a direction the paper itself flags.
  • The ablation suggests segmentation is the largest single contributor to detection quality (recall from 80.6% to 94.1%), implying that the segmentation branch does real work rather than acting as a regularizer; this could be tested on other small-object aerial detection datasets.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 7 minor

Summary. The manuscript proposes AeroLite-MDNet, a lightweight multi-task network combining a detection head and a segmentation head for detecting deviations during UAV landing, together with a new dataset (UAVLandData) of 9,142 images annotated with bounding boxes and segmentation masks, and a new evaluation metric (Average Warning Delay, AWD). The system center-crops each frame to 640x640, runs detection and segmentation, and uses Algorithm 1 to issue a warning when the target's offset from the image center exceeds a threshold. The paper reports 98.6% warning accuracy, 0.7 s AWD, 7.12M parameters, 104.2 FPS on an RTX 3070, and 7.3 FPS on a Jetson TX2, and compares against six baselines in detection, segmentation, and warning-system metrics.

Significance. If the results hold, the contribution is practically useful: the model is compact, the multi-task detection-segmentation design is reasonable, the UAVLandData dataset is a potentially valuable community resource, and the paper explicitly commits to releasing code. The novelty is incremental rather than foundational, but the application is timely. However, the headline real-time claim rests on an AWD definition that explicitly excludes inference latency, and the ground-truth deviation onset and the decision threshold are not defined; as written, the experimental support for 'effective real-time warning' is incomplete and needs substantial revision.

major comments (4)
  1. [Section III-D, Eq. (15); Section IV-C, Table IV; Conclusion] The AWD metric as defined does not measure wall-clock warning delay. The text states that 'frames captured during the model's processing time are excluded. Instead, the last frame captured before inference completes is used.' This means the warning time Pi in Eq. (15) is the capture timestamp of an input frame, not the time at which the warning is issued after inference. For a system that warns only after inference finishes, the true delay is at least Pi + k, where k is the per-frame processing time. On the Jetson TX2, k is about 0.137 s at 7.3 FPS, and for DeepLabV3+ at 0.2 FPS, k is about 5 s, so the omission can understate AWD by seconds and change the relative ranking of methods. The abstract's 'AWD of 0.7 seconds' and the conclusion's claim that the system 'meets the real-time warning requirements' are therefore not supported as stated. The authors should report AWD including inference latency, or explicitly redefine the metric as a frame-index perception lag and avoid claiming real-time wall-clock responsiveness based on it.
  2. [Section III-D, Eq. (15); Section III-C, Algorithm 1; Section IV-C] The ground-truth deviation onset time Ti in Eq. (15) is never defined. To compute AWD, the authors must specify how the 'actual deviation time' is determined for each landing sequence: is it manually annotated, is it derived from the true nest position crossing a threshold, and what is the protocol for the early phase when only houses are visible? Relatedly, the threshold delta in Algorithm 1 is never reported, yet ACC, FPR, and AWD all depend on it. Without this information, the central warning-system metrics are not reproducible. The paper should state the onset annotation protocol, the value of delta, and how the five landing stages in Fig. 6 are segmented in time.
  3. [Section IV-C, Tables III and IV] All performance numbers are reported as single-point estimates with no error bars, no number of repeated runs, and no statement of the number of landing sequences n used in Eq. (15). Given that the test split contains only 1,636 images and the warning-system metrics are aggregate over entire landing trajectories, the authors should report variance across runs or folds, at least for the headline metrics (AWD, ACC, FPR). They should also state how many landing videos/sequences contribute to Table IV.
  4. [Section IV-A and Fig. 2] The deviation rule in Algorithm 1 and the center-crop preprocessing both assume that the camera optical center is the true landing position and that the landing zone is centered in the field of view at the start of descent. If the camera is mounted with a known offset or the target is not centered initially, the labels and the AWD measure target-to-image-center alignment rather than actual landing deviation. The authors should validate this assumption (e.g., with camera calibration or by annotating the true touchdown point) and state what happens when the center-crop moves the target outside the field of view. A concrete test would be to compare warnings based on image-center distance with warnings based on a calibrated ground-truth landing point.
minor comments (7)
  1. [Abstract and Section I] The code availability URL in the abstract (https://github.com/ITTTTTI/Maskyolo.git) differs from the one in the first-page footnote (https://github.com/AeroLite-MDNet/AeroLite-MDNet.git); the two should be aligned.
  2. [Section III-B1, Eq. (10)] The CIoU formula appears malformed: the first term is written as rho^2(c2,c1) in both numerator and denominator, and the roles of alpha and v are not those of the standard CIoU loss. Please check the equation against the cited source and correct it.
  3. [Section III-B2, Eq. (14)] Equation (14) contains a dangling phrase 'where y_i is the true label' after an incomplete expression; the sentence 'where y_hat_i is the output, , and y_i is the true label' should be cleaned up.
  4. [Section III-B, Algorithm 2] Algorithm 2's loop condition 'while accuracy of verification set > 90%' is inconsistent with the text that says 'We stop training until the model achieves an accuracy greater than 90%'; the condition should be inverted or clarified.
  5. [Section III-B1] The focal-loss description states that 'When a sample is difficult to classify, p_t approaches 1', but for a difficult sample p_t is small, while for an easy sample p_t approaches 1 and the factor (1-p_t)^gamma reduces the loss weight. The sentence should be corrected.
  6. [Table III] The header and body of Table III are misaligned in the typeset text; for example, the UNet row renders the FPS TX2 and FPS GPU values as '0.5243.9'. Please reformat the table so each column is unambiguous.
  7. [Section IV-C, Table IV] The sentence 'our false alarm rate is slightly higher than Mask R-CNN and lower than ConvNeXt-V2' contradicts the table, which shows FPR of 0.7 for both Mask R-CNN and the proposed method and 0.4 for ConvNeXt-V2; the text should be corrected.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the paper's headline numbers are held-out empirical evaluations, and no prediction reduces to a fitted constant or to a self-citation chain.

full rationale

The central claims in this paper are empirical: AeroLite-MDNet's 98.6% warning accuracy, 0.7 s AWD, 7.12M parameters, 104.2 FPS on RTX 3070, and 7.3 FPS on Jetson TX2 are reported as measured outcomes on a held-out test split (Tables III and IV). No fitted parameter is renamed as a prediction: the anchor-box aspect ratios are standard hyperparameters chosen from labeled data statistics, the segmentation loss weights (alpha1=0.5, alpha2=0.25, alpha3=0.25) are tuning choices, and the reported accuracy is evaluated on data not used for training. Algorithm 1 defines the deviation warning rule from image-center geometry; this is the task definition, not a result derived from the model itself. Section III-D's AWD (Eq. 15) is a metric definition rather than a derived prediction; the decision to exclude frames captured during inference is a measurement-validity concern about whether AWD reflects wall-clock warning delay, not a circularity in the derivation. The paper contains no load-bearing self-citations: all architecture components (CA Conv, SCA C3, multi-scale segmentation head) are introduced directly and compared against external baselines. The center-crop assumption in Section IV-A is an external-validity limitation, not a circular step, because the model is not given the ground-truth deviation labels at inference time. Overall, the evaluation is self-contained and the claimed numbers stand or fall on the quality of the data and experimental protocol rather than on any by-construction equivalence.

Assumptions & free parameters 5 free parameters · 4 assumptions · 1 invented entities

The empirical claims rest on a dataset-specific set of anchor boxes, tuned loss weights, and an unreported deviation threshold. The decision rule assumes the image center is the true landing point, and the AWD metric is not anchored to an externally measured system delay.

free parameters (5)
  • Anchor box sizes = small (10,13),(16,30),(33,23); medium (30,61),(62,45),(59,119); large (116,90),(156,198),(373,326)
    Chosen from statistical analysis of labeled data (Section III-B1), so they are fitted to this dataset.
  • Segmentation loss weights = alpha1=0.5, alpha2=0.25, alpha3=0.25
    Selected 'after experimental testing' (Section III-B2), likely tuned on validation data.
  • Deviation threshold delta = not reported
    Algorithm 1 declares deviation when d/l > delta; the value is chosen by the authors but never stated, directly affecting ACC and AWD.
  • Learning rate and weight decay = 0.001 and 0.0005
    Training hyperparameters reported in Section IV-B; standard but still affect the final numbers.
  • Backbone C3 module depths = 3, 6, 9, 3
    Depth choices of the SCA_C3 modules in Fig. 2, manually selected for a lightweight trade-off.
assumptions (4)
  • standard math Standard deep learning loss and attention formulas from prior work are correct and applicable (Eqs. 1 through 14).
    Assumes focal loss, CIoU, Dice loss, and attention definitions as published.
  • domain assumption At the start of landing, the critical landing zone is centered in the camera view, so center-cropping to 640x640 preserves the target.
    Explicitly stated in Section IV-A; if false, the crop and the center-based deviation rule break.
  • domain assumption The image center is the actual landing position, and deviation can be measured by the distance from the detected target to that center.
    Algorithm 1 uses this rule to generate warning labels; no external landing accuracy measurement is used.
  • ad hoc to paper The active labeling strategy in Algorithm 2 selects informative samples by Euclidean distance in pixel space.
    Justification for this diversity heuristic is not provided; it is not used in the final evaluation but underlies dataset construction.
invented entities (1)
  • Average Warning Delay (AWD)
    purpose: New evaluation metric for warning latency.
    Defined by Eq. 15, but ground-truth onset times T_i are not defined and inference latency is excluded, so it lacks a clear external referent.

how reviews work

0 comments
Cite this review

Pith. "Pith review of AeroLite-MDNet: Lightweight Multi-task Deviation Detection Network for UAV Landing." pith.science (2026). https://pith.science/paper/7SYJ3HCP

@misc{pith2026250621635,
  author       = {Pith},
  title        = {Pith review of: AeroLite-MDNet: Lightweight Multi-task Deviation Detection Network for UAV Landing},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/7SYJ3HCP}},
  note         = {Machine review of arXiv:2506.21635}
}
read the original abstract

Unmanned aerial vehicles (UAVs) are increasingly employed in diverse applications such as land surveying, material transport, and environmental monitoring. Following missions like data collection or inspection, UAVs must land safely at docking stations for storage or recharging, which is an essential requirement for ensuring operational continuity. However, accurate landing remains challenging due to factors like GPS signal interference. To address this issue, we propose a deviation warning system for UAV landings, powered by a novel vision-based model called AeroLite-MDNet. This model integrates a multiscale fusion module for robust cross-scale object detection and incorporates a segmentation branch for efficient orientation estimation. We introduce a new evaluation metric, Average Warning Delay (AWD), to quantify the system's sensitivity to landing deviations. Furthermore, we contribute a new dataset, UAVLandData, which captures real-world landing deviation scenarios to support training and evaluation. Experimental results show that our system achieves an AWD of 0.7 seconds with a deviation detection accuracy of 98.6\%, demonstrating its effectiveness in enhancing UAV landing reliability. Code will be available at https://github.com/ITTTTTI/Maskyolo.git

Figures

Figures reproduced from arXiv: 2506.21635 by the authors.

Figure 1
Figure 1. Comparison of current methods and ours. (a) The multitask model [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. The proposed warning system includes seven steps: (1) during the UAV descent, the video stream is transmitted to the backend; (2) the input video [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Attention modules in the AeroLite-MDNet model, where (a) is the [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: Structure of segmentation head. For objects of different scales, we [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]
Figure 5
Figure 5. Figure 5: Representative images from the UAVLandData dataset. Normal: [PITH_FULL_IMAGE:figures/full_fig_p006_5.png]
Figure 6
Figure 6. Figure 6: Performance comparison of warning systems using different models during the five stages of descent (P1), stabilized descent (P2), approaching landing [PITH_FULL_IMAGE:figures/full_fig_p008_6.png]
Figure 7
Figure 7. Figure 7: System performance in practical applications: (a) Normal landing under adequate lighting conditions. (b) Normal landing under low-light conditions [PITH_FULL_IMAGE:figures/full_fig_p009_7.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

30 extracted references · 22 canonical work pages

  1. [1]

    The application of uav for the analysis of geological hazard in krk island, croatia, mediterranean sea,

    I. Ru ˇzi´c, ˇC. Benac, S. D. Jovan ˇcevi´c, and M. Radi ˇsi´c, “The application of uav for the analysis of geological hazard in krk island, croatia, mediterranean sea,”Remote Sensing, vol. 13, no. 9, p. 1790, 2021

    work page 2021

  2. [2]

    Survey of autonomous drone hangars–opportunities and challenges for maritime platforms,

    J. Knitter, D. Lai, L. Baxter, N. Jayarathne, and S. De Vincentis, “Survey of autonomous drone hangars–opportunities and challenges for maritime platforms,” inProceedings of the International Conference on, vol. 2, pp. 81–87, 2024

    work page 2024

  3. [3]

    Estimation techniques in robust vision-based landing of aerial vehicles,

    T. Nakamura, D. Magree, and E. N. Johnson, “Estimation techniques in robust vision-based landing of aerial vehicles,”IFAC-PapersOnLine, vol. 50, no. 1, pp. 11 664–11 669, 2017

    work page 2017

  4. [4]

    Consumer-grade global positioning system (gps) accuracy and reliability,

    M. G. Wing, A. Eklund, and L. D. Kellogg, “Consumer-grade global positioning system (gps) accuracy and reliability,”Journal of forestry, vol. 103, no. 4, pp. 169–173, 2005

    work page 2005

  5. [5]

    Dino: Detr with improved denoising anchor boxes for end-to- end object detection,

    H. Zhang, F. Li, S. Liu, L. Zhang, H. Su, J. Zhu, L. M. Ni, and H.-Y . Shum, “Dino: Detr with improved denoising anchor boxes for end-to- end object detection,” 2022

    2022

  6. [6]

    Autonomous landing for a multirotor uav using vision,

    S. Lange, N. S ¨underhauf, and P. Protzel, “Autonomous landing for a multirotor uav using vision,” inInternational conference on simulation, modeling, and programming for autonomous robots (SIMPAR 2008), pp. 482–491, 2008

    work page 2008

  7. [7]

    Implementation of vision-based real time helipad detection system,

    S. Lee, J.-w. Jang, and K.-R. Baek, “Implementation of vision-based real time helipad detection system,” in2012 12th International Conference on Control, Automation and Systems, pp. 191–194, 2012

    work page 2012

  8. [8]

    Embedded vision system for auto- mated drone landing site detection,

    P. Fraczek, A. Mora, and T. Kryjak, “Embedded vision system for auto- mated drone landing site detection,” inComputer Vision and Graphics: International Conference, ICCVG 2018, Warsaw, Poland, September 17- 19, 2018, Proceedings, pp. 397–409. Springer, 2018

    work page 2018

Show all 30 references
  1. [9]

    Autonomous landing of uav based on artificial neural network supervised by fuzzy logic,

    J. P. C. de Souza, A. L. M. Marcato, E. P. de Aguiar, M. A. Juc ´a, and A. M. Teixeira, “Autonomous landing of uav based on artificial neural network supervised by fuzzy logic,”Journal of Control, Automation and Electrical Systems, vol. 30, pp. 522–531, 2019

    2019

  2. [10]

    Vision- based uav guidance for autonomous landing with deep neural networks,

    Y . Bicer, M. Moghadam, C. Sahin, B. Eroglu, and N. K. ¨Ure, “Vision- based uav guidance for autonomous landing with deep neural networks,” inAIAA Scitech 2019 Forum, p. 0140, 2019

    2019

  3. [11]

    A real-time semantic segmentation method based on stdc-ct for recognizing uav emergency landing zones,

    B. Jiang, Z. Chen, J. Tan, R. Qu, C. Li, and Y . Li, “A real-time semantic segmentation method based on stdc-ct for recognizing uav emergency landing zones,”Sensors, vol. 23, no. 14, p. 6514, 2023

    2023

  4. [12]

    Swin-yolox for autonomous and accurate drone visual landing,

    R. Chen, Y . Xu, M. S. b. Sinal, D. Zhong, X. Li, B. Li, Y . Guo, and Q. Luo, “Swin-yolox for autonomous and accurate drone visual landing,” IET Image Processing, vol. 18, no. 14, pp. 4731–4744, 2024

    2024

  5. [13]

    Faster r-cnn: Towards real-time object detection with region proposal networks,

    S. Ren, K. He, R. Girshick, and J. Sun, “Faster r-cnn: Towards real-time object detection with region proposal networks,”Advances in neural information processing systems, vol. 28, 2015

    2015

  6. [14]

    Ssd: Single shot multibox detector,

    W. Liu, D. Anguelov, D. Erhan, C. Szegedy, S. Reed, C.-Y . Fu, and A. C. Berg, “Ssd: Single shot multibox detector,” inComputer Vision– ECCV 2016: 14th European Conference, Amsterdam, The Netherlands, October 11–14, 2016, Proceedings, Part I 14, pp. 21–37. Springer, 2016

    2016

  7. [15]

    Detrs with collaborative hybrid as- signments training,

    Z. Zong, G. Song, and Y . Liu, “Detrs with collaborative hybrid as- signments training,” inProceedings of the IEEE/CVF international conference on computer vision, pp. 6748–6758, 2023

    2023

  8. [16]

    ultralytics/yolov5: v6. 0-yolov5n’nano’models, roboflow integration, tensorflow export, opencv dnn support,

    G. Jocher, A. Stoken, A. Chaurasia, J. Borovec, Y . Kwon, K. Michael, L. Changyu, J. Fang, P. Skalski, A. Hoganet al., “ultralytics/yolov5: v6. 0-yolov5n’nano’models, roboflow integration, tensorflow export, opencv dnn support,”Zenodo, 2021

    2021

  9. [17]

    Fully convolutional networks for semantic segmentation,

    J. Long, E. Shelhamer, and T. Darrell, “Fully convolutional networks for semantic segmentation,” inProceedings of the IEEE conference on computer vision and pattern recognition, pp. 3431–3440, 2015

    2015

  10. [18]

    Segnet: A deep con- volutional encoder-decoder architecture for image segmentation,

    V . Badrinarayanan, A. Kendall, and R. Cipolla, “Segnet: A deep con- volutional encoder-decoder architecture for image segmentation,”IEEE transactions on pattern analysis and machine intelligence, vol. 39, no. 12, pp. 2481–2495, 2017

    2017

  11. [19]

    Segformer: Simple and efficient design for semantic segmentation with transformers,

    E. Xie, W. Wang, Z. Yu, A. Anandkumar, J. M. Alvarez, and P. Luo, “Segformer: Simple and efficient design for semantic segmentation with transformers,”Advances in neural information processing systems, vol. 34, pp. 12 077–12 090, 2021

    2021

  12. [20]

    Mask r-cnn,

    K. He, G. Gkioxari, P. Dollar, and R. Girshick, “Mask r-cnn,”2017 IEEE International Conference on Computer Vision (ICCV), Oct. 2017

    2017

  13. [21]

    Fast r-cnn,

    R. Girshick, “Fast r-cnn,” inProceedings of the IEEE international conference on computer vision, pp. 1440–1448, 2015

    2015

  14. [22]

    Rtmdet: An empirical study of designing real-time object detectors,

    C. Lyu, W. Zhang, H. Huang, Y . Zhou, Y . Wang, Y . Liu, S. Zhang, and K. Chen, “Rtmdet: An empirical study of designing real-time object detectors,” 2022

    2022

  15. [23]

    Focal loss for dense object detection,

    T.-Y . Lin, P. Goyal, R. Girshick, K. He, and P. Doll ´ar, “Focal loss for dense object detection,” inProceedings of the IEEE international conference on computer vision, pp. 2980–2988, 2017

    2017

  16. [24]

    Enhancing geometric factors in model learning and inference for object detection and instance segmentation,

    Z. Zheng, P. Wang, D. Ren, W. Liu, R. Ye, Q. Hu, and W. Zuo, “Enhancing geometric factors in model learning and inference for object detection and instance segmentation,”IEEE transactions on cybernetics, vol. 52, no. 8, pp. 8574–8586, 2021

    2021

  17. [25]

    V-net: Fully convolutional neural networks for volumetric medical image segmentation,

    F. Milletari, N. Navab, and S. A. Ahmadi, “V-net: Fully convolutional neural networks for volumetric medical image segmentation,”arXiv e- prints, 2016

    2016

  18. [26]

    U-net: Convolutional networks for biomedical image segmentation,

    O. Ronneberger, P. Fischer, and T. Brox, “U-net: Convolutional networks for biomedical image segmentation,” inMedical Image Computing and Computer-Assisted Intervention–MICCAI 2015: 18th International Conference, Munich, Germany, October 5-9, 2015, Proceedings, Part III 18, pp...

    2015

  19. [27]

    Encoder- decoder with atrous separable convolution for semantic image segmen- tation,

    L.-C. Chen, Y . Zhu, G. Papandreou, F. Schroff, and H. Adam, “Encoder- decoder with atrous separable convolution for semantic image segmen- tation,” inProceedings of the European conference on computer vision (ECCV), pp. 801–818, 2018

    2018

  20. [28]

    Mask r-cnn,

    K. He, G. Gkioxari, P. Doll ´ar, and R. Girshick, “Mask r-cnn,” in Proceedings of the IEEE international conference on computer vision, pp. 2961–2969, 2017

    2017

  21. [29]

    Convnext v2: Co-designing and scaling convnets with masked autoencoders,

    S. D. Sanghyun Woo, X. C. Ronghang Hu, I. S. K. Zhuang Liu, and S. Xie, “Convnext v2: Co-designing and scaling convnets with masked autoencoders,”arXiv preprint arXiv:2301.00808, 2023. Haiping Yangreceived the Ph.D. degree in car- tography and geographic information system fro...

    2023 arXiv

  22. [2015]

    Her research in- terests include high spatial resolution remote sensing image classification, change detection, and neural networks

    She is currently an Assistant Professor with Zhejiang University of Technology. Her research in- terests include high spatial resolution remote sensing image classification, change detection, and neural networks. Huaxing Liureceived the B.S. degree in ma- rine technology from ...

    2018

Pith tools

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