Pith. sign in

REVIEW 4 major objections 5 minor 41 references

TAME: Temporal Audio-based Mamba for Enhanced Drone Trajectory Estimation and Classification

T0 review · 4 major / 5 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read An audio-only model estimates drone trajectories and classifies drone types better than visual and fusion baselines.

desk verdict Audio-only drone tracker with a sensible Mamba fusion idea, but the headline SOTA claim is unverifiable as reported until the patch geometry, the day/night audio rows, and the missing error bars are sorted out. read the letter →

arxiv 2412.13037 v7 pith:ED3WAXKL submitted 2024-12-17 cs.SD eess.AS

Zhenyuan Xiao , Huanran Hu , Guili Xu , Junwei He This is my paper · ORCID
classification cs.SDeess.AS
keywords audio-baseddronedetectionstatespacemodelstrajectoryestimationclassificationmel-spectrogramcross-attentionanti-UAVtemporalfeatureenhancement
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

This paper introduces TAME, an audio-only system that estimates a drone's 3D trajectory and classifies the drone type from four channels of sound. The method turns multichannel audio into mel-spectrograms and runs parallel Mamba-style state-space scans along the time and frequency axes, then fuses the two streams with residual cross-attention. On the MMAUD V1 benchmark, the paper reports a mean trajectory error of 0.55 m and 98.0 percent classification accuracy, better than the published visual, audio, and audio-visual fusion methods it compares against. If those numbers hold, cheap passive microphones could serve as a standalone drone-threat sensor, including in conditions where cameras fail.

What carries the argument

The load-carrying object is TSMamba, a parallel pair of selective state-space models applied to a mel-spectrogram. TMamba scans the spectrogram from left to right along time with a learnable token, capturing temporal-difference-of-arrival features; SMamba scans top to bottom along frequency, capturing global spectral attenuation. The Temporal Feature Enhancement module then injects spectral information into temporal features via residual cross-attention, and the enhanced temporal token is read out by two MLP heads—one for 3D trajectory regression with L1 loss and one for drone-type classification with cross-entropy loss.

What would settle it

Run the released code on MMAUD V1 with the stated settings and compare the reproduced APE and accuracy with the paper's 0.55 and 98.0; also check whether the day and night audio test files differ. If the numbers do not reproduce, or the day and night audio are identical, the central claim fails.

Watch

Extended reading notes

Core claim

The central claim is that a temporal–spectral Mamba backbone, called TSMamba, can extract from audio alone the information needed to locate a drone in 3D and identify its model. The paper's mechanism is that a temporal scan of the mel-spectrogram encodes time-difference-of-arrival features that carry distance and, in part, category information, while a spectral scan encodes the sound-attenuation profile across frequencies. A Temporal Feature Enhancement module merges the spectral features into the temporal features using cross-attention with a residual connection, so that a single learnable temporal token carries the information used by two lightweight heads. TAME is reported to reach 0.55 average position error and 98.0 percent accuracy on MMAUD V1, outperforming the strongest fusion baseline AV-FDTI at 1.24 average position error and 94.0 percent accuracy.

Load-bearing premise

The claimed results depend on the experiments being run exactly as described, including genuinely separate day and night audio recordings; if the description and the actual run diverge, the comparison against other methods is not yet established.

Editorial extensions

If this is right

  • Audio-only detection could make drone monitoring systems smaller, cheaper, and harder to detect, since they need only a microphone array rather than cameras or radar.
  • The method's night-time numbers being equal to its day-time numbers implies the audio features are treated as illumination-independent; if that holds, audio-only systems would remain useful exactly where visual trackers fail.
  • Because the model is a generic temporal/spectral sequence learner, the same TSMamba-TFE design can be reapplied to other tasks that need location and category from one sensor stream.
  • Releasing code and trained weights lets any group rerun the MMAUD V1 comparison, so the claimed lead over fusion baselines is directly checkable.

Reading between the lines

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

  • The paper does not report how performance varies with distance or noise level; a natural next experiment is to measure trajectory error in decibel bins to see where audio-only estimation starts to fail.
  • A stronger test of the paper's 'especially at night' framing would record identical drone flights in true day and night conditions, since the current audio rows do not vary with illumination.
  • The architecture could be transferred to nearby problems such as speaker localization or acoustic event direction-finding, where the temporal-difference and spectral-attenuation features have the same physical meaning.
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 / 5 minor

Summary. The manuscript proposes TAME, an audio-only model combining a temporal-spectral Mamba (TSMamba) backbone with a temporal feature enhancement (TFE) cross-attention module, for 3D drone trajectory estimation and drone-type classification on the MMAUD dataset. The model ingests multichannel audio mel-spectrograms, splits them into temporal and spectral patch sequences, processes each with selective state-space models, fuses them, and feeds a learned token to two MLP heads. The paper reports state-of-the-art results on MMAUD V1, with an average APE of 0.55 and classification accuracy of 98.0%, outperforming visual, audio, and audio-visual fusion baselines, and includes an ablation study and a public code release.

Significance. If the reported results are reproducible, the paper is significant: it suggests that a purely audio-based model with a state-space backbone can match or exceed audio-visual fusion methods on a recent multi-modal drone benchmark, which would be practically relevant for low-cost and lighting-invariant drone detection. The architecture is clearly motivated and the open-source release is a concrete strength. The main caveat is that the evaluation as reported is not yet reproducible: the patch-splitting configuration is internally inconsistent, the experimental protocol omits several key details, and the headline comparison rests on single-run numbers without uncertainty estimates. These issues are fixable, but they currently prevent full verification of the central claim.

major comments (4)
  1. [§II-A, §III-B] The patch-splitting configuration is internally inconsistent. With a 224×16 mel-spectrogram and the stated J=16, W=4, H=1, the temporal patch sequence has R/W = 224/4 = 56 patches, not 16, while the spectral patch sequence has S/H = 16/1 = 16 patches. Equations (1) and (2) therefore cannot both use the same J for the two patch sequences at the stated input resolution. Please clarify which J applies to each stream, or correct the patch dimensions; this is load-bearing because the architecture shape and the reported results depend on the actual splitting scheme.
  2. [Table I, §III-C] The evaluation protocol does not support the strength of the stated conclusions. The table reports a single run per method with no error bars, multiple seeds, or statistical tests, so the headline improvement (TAME APE 0.55 vs. AV-FDTI 1.24) could be within run-to-run variation. In addition, the audio-only rows report identical Light and Dark values for TAME, AudioNet, and DroneChase; while this is plausible for a lighting-invariant audio model, the text's claim of superiority 'especially at night' is not substantiated by any night-specific audio result, and the paper does not state how the day/night split is defined for audio recordings. Please provide uncertainty estimates and clarify the day/night audio protocol.
  3. [§III-B] The term 'Brightness attenuation' is applied during training/testing for an audio-only model, but the augmentation is never defined for audio or mel-spectrograms. It is unclear whether this is a visual augmentation erroneously carried over from an image pipeline, or an audio-specific transformation, and no parameters are given. This ambiguity prevents reproduction of the reported experimental pipeline and should be resolved by either defining the audio-domain augmentation precisely or removing the reference.
  4. [§II-A, §III-C] The comparison omits Mamba-based audio baselines such as Audio Mamba [35] and SSAMBA [36], which are cited in the introduction but not evaluated. Since the contribution is specifically a Mamba-based audio architecture, including at least one such baseline would substantially strengthen the claim that the proposed temporal-spectral design, rather than the SSM backbone alone, is responsible for the gains. I regard this as a completeness issue that should be addressed in revision.
minor comments (5)
  1. [§III-A] The mel-spectrogram extraction is not fully specified; please state the STFT window size, hop length, number of mel filters, and whether the 224×16 dimensions correspond to time×frequency or frequency×time, as the patch-splitting description depends on this orientation.
  2. [Abstract] The dataset name is written as 'MMUAD' in the abstract but as 'MMAUD' elsewhere, including reference [41]; please unify the spelling.
  3. [Table I] The table header is difficult to parse; the column grouping under Light, Dark, and Overline should be made explicit, and the 'Params' column should be clearly associated with the whole row rather than the Modality column.
  4. [§II-A] The state-space model equations are not given; a brief description of the selective SSM update and the role of the learnable token would improve reproducibility, especially since the implementation is said to be public.
  5. [§IV] The conclusion mentions 'reliance on large datasets' as a limitation, but the paper does not discuss the size of the training set or how the 7:3 split was performed; please add this information to the experimental section.

Circularity Check

0 steps flagged · score 0.0 of 10

No circular derivation: the SOTA claim is an empirical benchmark comparison against the external MMAUD dataset, not a result derived from its own inputs.

full rationale

The paper's central claim is that TAME achieves state-of-the-art trajectory estimation and classification on the MMAUD benchmark. This is an empirical, externally checkable claim: the model is trained on audio mel-spectrograms from MMAUD and evaluated with APE and accuracy against published baselines. No parameter is fitted to a subset of the benchmark and then renamed as a prediction; no equation defines the output in terms of the input in a way that would make the result true by construction. The architecture (TSMamba, TFE, detection head) is described with explicit loss functions, and the reported gains are benchmark numbers, not derivations. The paper cites prior work from the same research community, including the MMAUD dataset and baselines, but those citations are used as external benchmarks and comparison methods, not as load-bearing justification for TAME's own design choices. Notably, the TAME authors do not overlap with the authors of the MMAUD dataset paper or the main baseline papers cited for comparison. There is a self-citation to the authors' related AV-DTEC work, but it is not used to justify the central claim. Serious internal inconsistencies exist—identical Light and Dark columns for audio-only rows including TAME, and a patch-splitting configuration (J=16, W=4, H=1) that does not match the stated 224x16 spectrogram—but these are reproducibility and verification concerns, not circularity. They should be assessed as correctness risks, not as evidence that the result reduces to its own inputs. Therefore, no significant circularity is present.

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

The central results depend on benchmark validity, hand-chosen architecture and training hyperparameters, and the unstated equivalence of audio-only evaluation across day and night conditions.

free parameters (4)
  • multi-task loss weight gamma = 2
    Set manually in Eq. (5) to balance classification and trajectory losses; no sensitivity analysis is reported.
  • model dimensions (L=12, n=6, dm=dk=192, J=16, W=4, H=1) = L=12, n=6, dm=dk=192, J=16, W=4, H=1
    Architecture hyperparameters from Section III-B; no ablation for these except module-level ablations in Table II.
  • training schedule (learning rate, batch size, epochs) = 1e-4, 64, 200
    Adam optimizer settings in Section III-B; chosen without a reported tuning procedure.
  • mel-spectrogram resolution and segment length = 224x16, 0.2 s
    Input representation choices from Section III-A; no analysis of sensitivity to these values.
assumptions (4)
  • standard math Selective state-space model (Mamba) and dot-product attention operations are valid and correctly implemented.
    The paper relies on Mamba [39] and dot-product attention in Eq. (2).
  • domain assumption The MMAUD benchmark's audio and ground-truth 3D trajectories are accurate and the 7:3 split is unbiased.
    Section III-A uses the MMAUD dataset and split; no labeling error analysis is provided.
  • domain assumption 0.2-second audio segments contain enough information for 3D trajectory estimation and classification.
    Input representation in Section III-A; no ablation on segment length is reported.
  • ad hoc to paper 'Brightness attenuation' is a meaningful augmentation for audio mel-spectrograms.
    Mentioned in Section III-B without explanation; likely inherited from visual pipelines.

how reviews work

0 comments
Cite this review

Pith. "Pith review of TAME: Temporal Audio-based Mamba for Enhanced Drone Trajectory Estimation and Classification." pith.science (2026). https://pith.science/paper/ED3WAXKL

@misc{pith2026241213037,
  author       = {Pith},
  title        = {Pith review of: TAME: Temporal Audio-based Mamba for Enhanced Drone Trajectory Estimation and Classification},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ED3WAXKL}},
  note         = {Machine review of arXiv:2412.13037}
}
read the original abstract

The increasing prevalence of compact UAVs has introduced significant risks to public safety, while traditional drone detection systems are often bulky and costly. To address these challenges, we present TAME, the Temporal Audio-based Mamba for Enhanced Drone Trajectory Estimation and Classification. This innovative anti-UAV detection model leverages a parallel selective state-space model to simultaneously capture and learn both the temporal and spectral features of audio, effectively analyzing propagation of sound. To further enhance temporal features, we introduce a Temporal Feature Enhancement Module, which integrates spectral features into temporal data using residual cross-attention. This enhanced temporal information is then employed for precise 3D trajectory estimation and classification. Our model sets a new standard of performance on the MMUAD benchmarks, demonstrating superior accuracy and effectiveness. The code and trained models are publicly available on GitHub https://github.com/AmazingDay1/TAME.

Figures

Figures reproduced from arXiv: 2412.13037 by the authors.

Figure 1
Figure 1. Motivation of our proposed solution. [22], radar, or images [23], but these approaches can be error-prone, particularly due to limitations in radar cross sec￾tion [24]–[26] or pixel size [27]. While some methods attempt to overcome these issues through multi-modal fusion [28]– [31], combining inputs like multi-spectral images and audio data, they often rely on unrealistic assumptions [32], such as perfect targeting … view at source ↗
Figure 2
Figure 2. Proposed TAME Architecture for audio-only UAV detection. [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗
Figure 3
Figure 3. Temporal Feature Enhancement Module height, W is the width of the patch, and H as the height of the patch. To extract audio features, multichannel audio is converted into mel-spectrograms χ ∈ R κ×R×S , which serve as input to our model. As the standard Mamba is a 1D sequence model, the spectrogram is split and flattened along the temporal and spectral axis to obtain p∗ , which consists of the temporal patch sequence… view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Test set trajectory estimation: Red curves represent ground truth, blue curves show predicted trajectories. [PITH_FULL_IMAGE:figures/full_fig_p004_4.png]
Figure 5
Figure 5. Figure 5: The confusion matrix for the classification results. [PITH_FULL_IMAGE:figures/full_fig_p004_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

41 extracted references · 33 canonical work pages

  1. [35]

    Audio mamba: Bidirectional state space model for audio representa- tion learning,

    Mehmet Hamza Erol, Arda Senocak, Jiu Feng, and Joon Son Chung, “Audio mamba: Bidirectional state space model for audio representa- tion learning,” arXiv preprint arXiv:2406.03344 , 2024

    arXiv 2024

  2. [36]

    Ssamba: Self-supervised audio representation learning with mamba state space model,

    Siavash Shams, Sukru Samet Dindar, Xilin Jiang, and Nima Mesgarani, “Ssamba: Self-supervised audio representation learning with mamba state space model,” arXiv preprint arXiv:2405.11831 , 2024

    arXiv 2024

  3. [1]

    Towards utilizing deep uncertainty in traditional slam,

    Mahdi Abolfazli Esfahani, Keyu Wu, Shenghai Yuan, and Han Wang, “Towards utilizing deep uncertainty in traditional slam,” in ICCA. IEEE, 2019

    work page 2019

  4. [2]

    Neptune: nonentangling trajectory planning for multiple tethered unmanned vehicles,

    Muqing Cao, Kun Cao, Shenghai Yuan, Thien-Minh Nguyen, and Lihua Xie, “Neptune: nonentangling trajectory planning for multiple tethered unmanned vehicles,” IEEE TRO, 2023

    work page 2023

  5. [3]

    Path planning for multiple tethered robots using topological braids,

    Muqing Cao, Kun Cao, Shenghai Yuan, Kangcheng Liu, Yan Loi Wong, and Lihua Xie, “Path planning for multiple tethered robots using topological braids,” RSS, 2023

    work page 2023

  6. [4]

    Distance-based multiple noncooperative ground target encirclement for complex environments,

    Fen Liu, Shenghai Yuan, Kun Cao, Wei Meng, and Lihua Xie, “Distance-based multiple noncooperative ground target encirclement for complex environments,” TCST, 2024

    work page 2024

  7. [5]

    A cost-effective cooperative exploration and inspection strategy for heterogeneous aerial system,

    Xinhang Xu, Muqing Cao, Shenghai Yuan, Thien Hoang Nguyen, Thien-Minh Nguyen, and Lihua Xie, “A cost-effective cooperative exploration and inspection strategy for heterogeneous aerial system,” in Proc. ICCA, 2024

    work page 2024

  8. [6]

    Non- cooperative stochastic target encirclement by anti-synchronization con- trol via range-only measurement,

    Fen Liu, Shenghai Yuan, Wei Meng, Rong Su, and Lihua Xie, “Non- cooperative stochastic target encirclement by anti-synchronization con- trol via range-only measurement,” in Proc. IEEE ICRA , 2023

    work page 2023

Show all 41 references
  1. [7]

    Spins: A structure priors aided inertial navigation system,

    Yang Lyu, Thien-Minh Nguyen, Liu Liu, Muqing Cao, Shenghai Yuan, Thien Hoang Nguyen, and Lihua Xie, “Spins: A structure priors aided inertial navigation system,” JFR, 2023

    2023

  2. [8]

    Survey on localization systems and algorithms for unmanned systems,

    Shenghai Yuan, Han Wang, and Lihua Xie, “Survey on localization systems and algorithms for unmanned systems,” Unmanned Systems, vol. 9, no. 02, pp. 129–163, 2021

    2021

  3. [9]

    Distributed multi-robot sweep coverage for a region with unknown workload distribution,

    Muqing Cao, Kun Cao, Xiuxian Li, Shenghai Yuan, Yang Lyu, Thien- Minh Nguyen, and Lihua Xie, “Distributed multi-robot sweep coverage for a region with unknown workload distribution,” Autonomous Intelligent Systems, 2021

    2021

  4. [11]

    Multi-robot active graph exploration with reduced pose-slam uncertainty via submodular optimization,

    Ruofei Bai, Shenghai Yuan, Hongliang Guo, Pengyu Yin, Wei-Yun Yau, and Lihua Xie, “Multi-robot active graph exploration with reduced pose-slam uncertainty via submodular optimization,” in IROS, 2024

    2024

  5. [12]

    Jacquard v2: Refining datasets using the human in the loop data correction method,

    Qiuhao Li and Shenghai Yuan, “Jacquard v2: Refining datasets using the human in the loop data correction method,” in ICRA, 2024

    2024

  6. [13]

    Direct: A differential dynamic programming based framework for trajectory generation,

    Kun Cao, Muqing Cao, Shenghai Yuan, and Lihua Xie, “Direct: A differential dynamic programming based framework for trajectory generation,” IEEE RAL, 2022

    2022

  7. [14]

    The 2nd anti-uav workshop & challenge: Methods and results,

    J Zhao, G Wang, J Li, L Jin, N Fan, M Wang, X Wang, T Yong, Y Deng, Y Guo, et al., “The 2nd anti-uav workshop & challenge: Methods and results,” arXiv preprint arXiv:2108.09909 , 2021

    2021 arXiv

  8. [15]

    The 3rd anti-uav workshop & challenge: Methods and results,

    Jian Zhao, Jianan Li, Lei Jin, Jiaming Chu, Zhihao Zhang, Jun Wang, Jiangqiang Xia, Kai Wang, Yang Liu, Sadaf Gulshad, et al., “The 3rd anti-uav workshop & challenge: Methods and results,” arXiv preprint arXiv:2305.07290, 2023

    2023 arXiv

  9. [16]

    Investigation of uav detection in images with complex backgrounds and rainy artifacts,

    Adnan Munir, Abdul Jabbar Siddiqui, and Saeed Anwar, “Investigation of uav detection in images with complex backgrounds and rainy artifacts,” in Proc. IEEE/CVF WACV, 2024

    2024

  10. [17]

    Air-to-air visual detection of micro-uavs: An experi- mental evaluation of deep learning,

    Ye Zheng, Zhang Chen, Dailin Lv, Zhixing Li, Zhenzhong Lan, and Shiyu Zhao, “Air-to-air visual detection of micro-uavs: An experi- mental evaluation of deep learning,” IEEE Robotics and automation letters, 2021

    2021

  11. [18]

    Unsupervised uav 3d trajectories estimation with sparse point clouds,

    Hanfang Liang, Yizhuo Yang, Jinming Hu, Jianfei Yang, Fen Liu, and Shenghai Yuan, “Unsupervised uav 3d trajectories estimation with sparse point clouds,” Proc. IEEE ICASSP , 2025

    2025

  12. [19]

    Audio array-based 3d uav trajectory estimation with lidar pseudo-labeling,

    Allen Lei, Tianchen Deng, Han Wang, Jianfei Yang, and Shenghai Yuan, “Audio array-based 3d uav trajectory estimation with lidar pseudo-labeling,” Proc ICASSP, 2025

    2025

  13. [20]

    Sepa- rating drone point clouds from complex backgrounds by cluster filter – technical report for cvpr 2024 ug2 challenge,

    Hanfang Liang, Jinming Hu, Xiaohuan Ling, and Bing Wang, “Sepa- rating drone point clouds from complex backgrounds by cluster filter – technical report for cvpr 2024 ug2 challenge,” arXiv preprint arXiv:2412.16947, 2024

    2024 arXiv

  14. [21]

    Large-scale uwb anchor calibration and one-shot localization using gaussian process,

    Shenghai Yuan, Boyang Lou, Thien-Minh Nguyen, Pengyu Yin, Muqing Cao, Xinghang Xu, Jianping Li, Jie Xu, Siyu Chen, and Lihua Xie, “Large-scale uwb anchor calibration and one-shot localization using gaussian process,” in arXiv preprint arXiv:2412.16880 , 2024

    2024 arXiv

  15. [22]

    Uloc: Learning to localize in complex large- scale environments with ultra-wideband ranges,

    Thien-Minh Nguyen, Yizhuo Yang, Tien-Dat Nguyen, Shenghai Yuan, and Lihua Xie, “Uloc: Learning to localize in complex large- scale environments with ultra-wideband ranges,” in arXiv preprint arXiv:2409.11122, 2024

    2024 arXiv

  16. [23]

    Autonomous object level segmenta- tion,

    Shenghai Yuan and Han Wang, “Autonomous object level segmenta- tion,” in Proc. ICARCV, 2014

    2014

  17. [24]

    Multi-modal uav detection, classification and tracking algorithm– technical report for cvpr 2024 ug2 challenge,

    Tianchen Deng, Yi Zhou, Wenhua Wu, Mingrui Li, Jingwei Huang, Shuhong Liu, Yanzeng Song, Hao Zuo, Yanbo Wang, Yutao Yue, et al., “Multi-modal uav detection, classification and tracking algorithm– technical report for cvpr 2024 ug2 challenge,” arXiv preprint arXiv:2405.16464, 2024

    2024 arXiv

  18. [25]

    A vehicle-mounted radar-vision system for precisely positioning cluster- ing uavs,

    Guangyu Wu, Fuhui Zhou, Kai Kit Wong, and Xiang-Yang Li, “A vehicle-mounted radar-vision system for precisely positioning cluster- ing uavs,” IEEE JSAC, 2024

    2024

  19. [26]

    On onboard lidar- based flying object detection,

    Matou ˇs Vrba, Viktor Walter, V´aclav Pritzl, Michal Pliska, Tom´aˇs B´aˇca, V ojtˇech Spurn `y, Daniel He ˇrt, and Martin Saska, “On onboard lidar- based flying object detection,” TRO, 2024

    2024

  20. [27]

    A real-time and lightweight method for tiny airborne object detection,

    Yanyi Lyu, Zhunga Liu, Huandong Li, Dongxiu Guo, and Yimin Fu, “A real-time and lightweight method for tiny airborne object detection,” in CVPR, 2023

    2023

  21. [28]

    Dronechase: A mobile and automated cross-modality system for continuous drone tracking,

    Neel V ora, Yi Wu, Jian Liu, and Phuc Nguyen, “Dronechase: A mobile and automated cross-modality system for continuous drone tracking,” in Proc. 9th MAVNetSysApp, 2023

    2023

  22. [29]

    Av- fdti: Audio-visual fusion for drone threat identification,

    Yizhuo Yang, Shenghai Yuan, Jianfei Yang, Thien Hoang Nguyen, Muqing Cao, Thien-Minh Nguyen, Han Wang, and Lihua Xie, “Av- fdti: Audio-visual fusion for drone threat identification,” Journal of Automation and Intelligence , 2024

    2024

  23. [30]

    Av-dtec: Self-supervised audio-visual fusion for drone trajectory estimation and classification,

    Zhenyuan Xiao, Yizhuo Yang, Guili Xu, Xianglong Zeng, and Shenghai Yuan, “Av-dtec: Self-supervised audio-visual fusion for drone trajectory estimation and classification,” arXiv preprint arXiv:2412.16928, 2024

    2024 arXiv

  24. [31]

    Mopa: Multi-modal prior aided domain adaptation for 3d semantic segmentation,

    Haozhi Cao, Yuecong Xu, Jianfei Yang, Pengyu Yin, Shenghai Yuan, and Lihua Xie, “Mopa: Multi-modal prior aided domain adaptation for 3d semantic segmentation,” in ICRA. IEEE, 2024

    2024

  25. [32]

    Anti-uav410: A thermal infrared benchmark and customized scheme for tracking drones in the wild,

    Bo Huang, Jianan Li, Junjie Chen, Gang Wang, Jian Zhao, and Tingfa Xu, “Anti-uav410: A thermal infrared benchmark and customized scheme for tracking drones in the wild,” IEEE TPAMI, 2023

    2023

  26. [33]

    A large-scale uav audio dataset and audio-based uav classification using cnn,

    Yaqin Wang, Zhiwei Chu, Ilmun Ku, E Cho Smith, and Eric T Matson, “A large-scale uav audio dataset and audio-based uav classification using cnn,” in Proc. IEEE IRC , 2022

    2022

  27. [34]

    Exploiting wavelet scattering transform & 1d-cnn for unmanned aerial vehicle detection,

    Murtiza Ali and Karan Nathwani, “Exploiting wavelet scattering transform & 1d-cnn for unmanned aerial vehicle detection,” IEEE Signal Processing Letters , 2024

    2024

  28. [37]

    Av-pedaware: Self-supervised audio-visual fusion for dynamic pedestrian awareness,

    Yizhuo Yang, Shenghai Yuan, Muqing Cao, Jianfei Yang, and Lihua Xie, “Av-pedaware: Self-supervised audio-visual fusion for dynamic pedestrian awareness,” in IROS, 2023

    2023

  29. [38]

    Is someone speaking? exploring long-term temporal features for audio-visual active speaker detection,

    Ruijie Tao, Zexu Pan, Rohan Kumar Das, Xinyuan Qian, Mike Zheng Shou, and Haizhou Li, “Is someone speaking? exploring long-term temporal features for audio-visual active speaker detection,” in Proc. ACM Int. Conf. Multimedia , 2021

    2021

  30. [39]

    Mamba: Linear-time sequence modeling with selective state spaces,

    Albert Gu and Tri Dao, “Mamba: Linear-time sequence modeling with selective state spaces,” arXiv preprint arXiv:2312.00752 , 2023

    2023 arXiv

  31. [40]

    An image is worth 16x16 words: Transformers for image recognition at scale,

    Alexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weis- senborn, Xiaohua Zhai, Thomas Unterthiner, Mostafa Dehghani, Matthias Minderer, Georg Heigold, Sylvain Gelly, et al., “An image is worth 16x16 words: Transformers for image recognition at scale,” ICLR, 2021

    2021

  32. [41]

    MMAUD: A Comprehensive Multi-Modal Anti-UA V Dataset for Modern Miniature Drone Threats,

    Shenghai Yuan, Yizhuo Yang, Thien Hoang Nguyen, Thien-Minh Nguyen, Jianfei Yang, Fen Liu, Jianping Li, Han Wang, and Lihua Xie, “MMAUD: A Comprehensive Multi-Modal Anti-UA V Dataset for Modern Miniature Drone Threats,” in ICRA, 2024

    2024

  33. [42]

    Yolov4: Optimal speed and accuracy of object detection,

    Alexey Bochkovskiy, Chien-Yao Wang, and Hong-Yuan Mark Liao, “Yolov4: Optimal speed and accuracy of object detection,” arXiv preprint arXiv:2004.10934, 2020

    2004 arXiv

Pith tools

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