Pith. sign in

REVIEW 4 major objections 4 minor 66 references

LASFNet: A Lightweight Attention-Guided Self-Modulation Feature Fusion Network for Multimodal Object Detection

T0 review · 4 major / 4 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 The paper claims that a single feature-level fusion unit, guided by attention and self-modulation, is enough to match or beat much heavier multimodal RGB-infrared detectors.

desk verdict A competent lightweight fusion paper whose most valuable contribution is the single-fusion-unit baseline, but the headline efficiency-accuracy triple is a cherry-picked composite and the reproducibility details need work. read the letter →

arxiv 2506.21018 v1 pith:AHMOVO6O submitted 2025-06-26 cs.CV

Lei Hao , Lina Xu , Chang Liu , Yanni Dong This is my paper
classification cs.CV
keywords multimodalobjectdetectionRGB-infraredfusionfeature-levellightweightnetworkattention-guidedself-modulationedgedeployment
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 claims that multimodal RGB-infrared object detection does not need multiple stacked feature-fusion units to be accurate; a single fusion stage followed by a lightweight attention module is enough. If true, the result matters because existing fusion-heavy detectors are too large and slow for drones, cars, and edge cameras. The authors build LASFNet on this baseline and report that it cuts parameter count by up to 90% and compute by up to 85% while raising detection mAP by 1–3 percent over state-of-the-art methods on DroneVehicle, LLVIP, and VTUAVdet. The paper is arguing for a simpler architecture, not just for a better score.

What carries the argument

The ASFF (attention-guided self-modulation feature fusion) module is the core unit: a three-stage process that performs channel and positional attention fusion, global and local feature modulation, and channel shuffle to fuse the RGB and infrared feature maps once in the backbone. A second module, FATM (feature attention transformation module), applies lightweight channel and positional attention at the neck to suppress redundant information and preserve multi-scale features. Together they let the network fuse modalities in a single pass instead of repeatedly across layers, which is what keeps the model light.

What would settle it

Run LASFNet and all compared detectors on the same three datasets at one fixed input resolution and one shared training schedule, then recompute mAP, parameters, and GFLOPs; if LASFNet no longer shows the 1–3 percent mAP advantage or the 85 percent GFLOPs reduction, the central efficiency-accuracy claim would fail.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central discovery is that a single feature-level fusion unit can match the accuracy of the popular multi-unit fusion baseline while removing roughly 36% of parameters and 20% of GFLOPs, and that the two proposed modules—ASFF and FATM—turn that slim baseline into a competitive detector. Across three datasets, LASFNet reports mAP improvements of 1–3 percent over state-of-the-art methods, with 7.7M parameters and 26.6 GFLOPs, compared with models ranging from 4.85M to 287.6M parameters and from 27.8 to 1038.2 GFLOPs. The argument is that attention-guided global and local self-modulation plus a neck-level attention transform recovers the representational capacity that multiple fusion units used to provide, at a fraction of the cost. This claim stands or falls on the fairness of the efficiency comparison and on whether the mAP gains reproduce under matched conditions.

Load-bearing premise

The efficiency and accuracy comparison assumes that every method was run at the same input resolution and under the same training protocol, but the paper states no input image size, so the reported FLOPs reductions and mAP margins may not hold under perfectly matched settings.

Editorial extensions

If this is right

  • A single fusion unit can serve as a default starting point for lightweight multimodal detectors, simplifying both training and deployment.
  • Models with far fewer parameters and FLOPs than current fusion-heavy detectors can still hold or improve accuracy on aligned and misaligned RGB-infrared benchmarks.
  • Isolating fusion into one stage makes it easier to swap backbones or detection heads without rebuilding the whole fusion stack.
  • The reported 7.7M-parameter and 26.6-GFLOP footprint fits embedded use cases such as drone controllers, autonomous driving modules, and smart cameras.
  • The gain from FATM suggests that neck-level attention is a cheap way to reduce information loss in feature pyramids and improve multi-scale detection.

Reading between the lines

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

  • If the single-fusion baseline holds, it suggests that prior multimodal works may have over-invested in fusion complexity, and the same simplification could transfer to other dual-modality tasks such as semantic segmentation or visual tracking.
  • A stronger test would be to compare the ASFF and FATM modules under a fixed compute budget, since the ablations show that adding them also increases parameters and GFLOPs; a budget-matched comparison would separate genuine fusion gains from extra capacity.
  • The claimed 85% compute reduction depends on input resolution, which the paper does not state, so readers should recompute FLOPs at an identical resolution before relying on the efficiency numbers.
  • The baseline claim can be tested directly by replacing ASFF with a single element-wise addition and checking whether the accuracy gap stays within the 1–2 mAP points reported in the paper.
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 / 4 minor

Summary. The paper proposes LASFNet, a YOLOv5-based dual-stream RGB-infrared detector for multimodal object detection. The architecture uses a single feature-level fusion unit (ASFF) in the backbone, which performs attention fusion, feature modulation fusion, and channel shuffle fusion, plus a lightweight Feature Attention Transformation Module (FATM) in the neck. Experiments are reported on DroneVehicle, LLVIP, and VTUAVdet, with ablation studies on LLVIP. The central claim is that LASFNet simultaneously cuts parameters by up to 90%, FLOPs by up to 85%, and improves mAP by 1%-3% relative to state-of-the-art methods.

Significance. The single-feature-fusion baseline is a legitimate and potentially useful design idea: Table I reports a 36% parameter reduction and 20% FLOPs reduction with no loss in mAP on LLVIP, and the ablation tables are directionally consistent with the proposed modules contributing to accuracy. The promise to release code is a strength. However, the empirical support is weakened by missing experimental protocol details, a baseline inconsistency between Tables II and III, and overstated reporting of improvement ranges. If these issues are fixed, the lightweight-fusion direction would be of interest to the multimodal detection community.

major comments (4)
  1. [Section IV-D, Tables II and III] The same baseline configuration is reported with different mAP values in Tables II and III. Table II, row 1 (Att=✕, FMB=✕, CS=✕) and Table III, row 1 (ASFF=✕, FATM=✕) are both described in Sections IV-D2 and IV-D3 as the single-fusion baseline using element-wise addition, with identical Params (7.2 M) and GFLOPs (20.8), yet the mAP values are 0.644 and 0.652. Since no seeds or repeated runs are reported, this inconsistency means that ablation deltas such as +0.002 for Att and +0.002 for CS in Table II cannot be distinguished from run-to-run variation. Please report means and standard deviations over multiple runs, or at least explain why the two baseline tables differ.
  2. [Section IV-B and Tables IV-VII] No input image resolution is stated for any dataset. Because GFLOPs scale with input resolution, the cross-method efficiency comparisons in Tables IV-VII are not verifiable unless all methods are evaluated at the same input size. Section IV-B gives training hyperparameters only for DroneVehicle (batch size 48, learning rate 0.01, 300 epochs); training settings for LLVIP and VTUAVdet are omitted. Please state the input resolution for each dataset and provide the full training and inference protocol for all three datasets.
  3. [Section IV-E] The sentence claiming a 'notable improvement of 2-5 percentage points' against SuperYOLO and GHOST 'across all three datasets' is not supported by the tables. On LLVIP, LASFNet improves over SuperYOLO by 9.4 percentage points (0.582 to 0.676) and over GHOST by 9.0 percentage points (0.586 to 0.676); on VTUAVdet the gain over GHOST is 21.0 percentage points (0.168 to 0.378). Please correct this statement or restrict it to the specific comparisons for which it actually holds.
  4. [Abstract and Section IV-E] The abstract's three-part efficiency-accuracy claim ('up to 90% parameters, 85% FLOPs, 1%-3% mAP') is not tied to any single comparison. The only row that simultaneously satisfies all three bounds is VTUAVdet versus GM-DETR in Table VI (89% parameter reduction, 84.9% FLOPs reduction, +2.8 percentage points mAP). Against GM-DETR on DroneVehicle the mAP gain is +8.8 percentage points (Table IV), on LLVIP it is +0.6 percentage points (Table V), and against CDC-YOLOFusion the FLOPs reduction is only about 4% (26.6 vs 27.8 GFLOPs) with mAP gains of +1.1 to +2.1 percentage points. Please specify exactly which baseline and dataset support each advertised number, and clarify whether '1%-3%' means absolute percentage points or relative improvement.
minor comments (4)
  1. [Equation (11)] There is an unbalanced parenthesis in the expression for F_m: 'Conv(Concat[φ(W(F_1)), F_2]))' has an extra closing parenthesis. Please correct the typo.
  2. [Section IV-C] The metric 'mAP' is not defined precisely; the paper should state the IoU threshold range (for example, COCO-style AP over 0.5:0.95) and how it differs from mAP50.
  3. [Section IV-B and Table VII] Table VII reports testing time on 'a single GPU,' but the GPU model is not specified, even though training is said to use two RTX 3090 GPUs. Please state the exact GPU and whether the reported time includes preprocessing and postprocessing.
  4. [Section III-B, Equation (6)] The text states that X_m has shape H×W×C, but the formula involving X_s ∈ R^{H/2×W/2×C} and σ²(X) ∈ R^{1×1×C} yields H/2×W/2×C before the upsampling operation in Equation (7). Please clarify the intermediate shapes and the broadcasting behavior.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity; the central efficiency-accuracy claim is an empirical benchmark comparison on held-out test sets, with only non-load-bearing self-citations.

full rationale

LASFNet's central efficiency-accuracy claim rests on Tables IV-VI, which compare measured parameters, GFLOPs, and mAP against external baselines on public benchmarks. These numbers are not derived from fitted constants, and no equation in the paper defines the reported mAP or efficiency in terms of the target claim. The proposed single-feature fusion baseline is validated by an independent ablation (Table I) against a multi-feature fusion baseline, and the ASFF/FATM ablations (Tables II-III) are component comparisons rather than circular definitions. Self-citations such as references [4], [9], [11], [46], and [47] appear in the related-work survey and do not supply any load-bearing premise. The abstract's simultaneous '90%/85%/1-3%' claim is not jointly supported by a single row of the tables (for example, the 85% GFLOPs reduction comes from comparing with GM-DETR, whose mAP gain on DroneVehicle is +8.8 points, not 1-3%), but this is a consistency and reproducibility concern, not a reduction of the result to its own inputs. No fitted parameter is relabeled as a prediction, no uniqueness theorem is imported from the authors' prior work, and no target result is assumed by construction. The paper is self-contained against external benchmarks, so the appropriate finding is no significant circularity; the score reflects only minor, non-load-bearing self-citations.

Assumptions & free parameters 4 free parameters · 5 assumptions · 0 invented entities

The paper introduces no new physical entities or forces. The only new objects are neural network modules (ASFF, FATM), which are combinations of known attention and self-modulation operations. The ledger lists the hyperparameters and modeling assumptions the empirical claims rest on.

free parameters (4)
  • alpha and beta (DFM modulation weights)
    Learnable scalars in Eq. (6) that balance variance versus downsampled global features. They are trained on data, and their initial values are not reported.
  • channel groups G in channel shuffle = not reported (likely 2)
    Hyperparameter in Eq. (13). No ablation over G is provided, and the text does not state the default value.
  • channel split ratio (C/2, 3C/2) in FM
    Split ratio chosen by hand in Eq. (11). No ablation justifies this particular partition.
  • loss weights lambda_box, lambda_obj, lambda_cls = YOLOv5 defaults (not stated)
    Eq. (19) uses three weighting hyperparameters. The paper does not specify their values or whether they were retuned per dataset.
assumptions (5)
  • domain assumption Dual-stream CSPDarknet53 from YOLOv5 provides suitable per-modality features for fusion.
    Section IV-B states the backbone is built upon CSPDarknet53, but no analysis of feature quality or pretraining is given.
  • domain assumption A single feature-level fusion unit at one scale is sufficient to propagate multimodal information to all detection scales.
    Table I supports this only for LLVIP with element-wise addition. The full LASFNet places ASFF at one location without searching over fusion depth or scale.
  • domain assumption Self-modulation blocks (DFM/FM) transfer from image super-resolution (SMFANet) to object detection without modification.
    The FMB is directly reused from a super-resolution context. No theoretical or empirical justification for the transfer is given beyond the reported ablations.
  • domain assumption All compared methods use the same input resolution, so GFLOPs and Params comparisons are fair.
    Implementation details do not report input sizes for any dataset, and FLOPs scale with resolution, making the efficiency comparison unverifiable.
  • domain assumption Dataset labels, splits, and evaluation protocols from prior works are accepted as ground truth.
    The paper reuses DroneVehicle, LLVIP, and VTUAVdet without verifying annotation quality, except a note in Figure 6 about LLVIP annotation omissions.

how reviews work

0 comments
Cite this review

Pith. "Pith review of LASFNet: A Lightweight Attention-Guided Self-Modulation Feature Fusion Network for Multimodal Object Detection." pith.science (2026). https://pith.science/paper/AHMOVO6O

@misc{pith2026250621018,
  author       = {Pith},
  title        = {Pith review of: LASFNet: A Lightweight Attention-Guided Self-Modulation Feature Fusion Network for Multimodal Object Detection},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/AHMOVO6O}},
  note         = {Machine review of arXiv:2506.21018}
}
read the original abstract

Effective deep feature extraction via feature-level fusion is crucial for multimodal object detection. However, previous studies often involve complex training processes that integrate modality-specific features by stacking multiple feature-level fusion units, leading to significant computational overhead. To address this issue, we propose a new fusion detection baseline that uses a single feature-level fusion unit to enable high-performance detection, thereby simplifying the training process. Based on this approach, we propose a lightweight attention-guided self-modulation feature fusion network (LASFNet), which introduces a novel attention-guided self-modulation feature fusion (ASFF) module that adaptively adjusts the responses of fusion features at both global and local levels based on attention information from different modalities, thereby promoting comprehensive and enriched feature generation. Additionally, a lightweight feature attention transformation module (FATM) is designed at the neck of LASFNet to enhance the focus on fused features and minimize information loss. Extensive experiments on three representative datasets demonstrate that, compared to state-of-the-art methods, our approach achieves a favorable efficiency-accuracy trade-off, reducing the number of parameters and computational cost by as much as 90% and 85%, respectively, while improving detection accuracy (mAP) by 1%-3%. The code will be open-sourced at https://github.com/leileilei2000/LASFNet.

Figures

Figures reproduced from arXiv: 2506.21018 by the authors.

Figure 1
Figure 1. Popular multimodal fusion baseline and the proposed new baseline. (a) Popular baseline with multi-feature fusion units: This approach requires a [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. The overall architecture of LASFNet. The detection network consists of three main components: the feature extraction backbone, the neck, and the [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Detailed architecture of our proposed attention-guided self-modulation feature fusion (ASFF) module. The fusion process is divided into three stages: [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: Detailed architecture of our proposed Feature Attention Transforma [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]
Figure 5
Figure 5. Figure 5: Three examples on DroneVehicle dataset to illustrate the detection results, where the purple triangles represent false positive objects, and the yellow [PITH_FULL_IMAGE:figures/full_fig_p010_5.png]
Figure 6
Figure 6. Figure 6: Three examples on LLVIP dataset to illustrate the detection results, where the purple triangles represent false positive objects, and the yellow triangles [PITH_FULL_IMAGE:figures/full_fig_p010_6.png]
Figure 7
Figure 7. Figure 7: Three examples on VTUAV det dataset to illustrate the detection results, where the purple triangles represent false positive objects, and the yellow triangles indicate the missing objects. TABLE V PERFORMANCE COMPARISON BETWEEN LASFNET AND SEVERAL STATE-OF-THE-ART APPR…

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

66 extracted references · 57 canonical work pages

  1. [1]

    Object detection in 20 years: A survey,

    Z. Zou, K. Chen, Z. Shi, Y . Guo, and J. Ye, “Object detection in 20 years: A survey,”Proc. IEEE, vol. 111, no. 3, pp. 257–276, 2023

    work page 2023

  2. [2]

    Ir reasoner: Real-time infrared object detection by visual reasoning,

    M. M. G ¨undo˘gan, T. Aksoy, A. Temizel, and U. Halici, “Ir reasoner: Real-time infrared object detection by visual reasoning,” inProc. CVPR, pp. 422–430, June 2023

    work page 2023

  3. [3]

    Camoformer: Masked separable attention for camouflaged object detection,

    B. Yin, X. Zhang, D.-P. Fan, S. Jiao, M.-M. Cheng, L. Van Gool, and Q. Hou, “Camoformer: Masked separable attention for camouflaged object detection,”IEEE Trans. Pattern Anal. Mach. Intell., vol. 46, pp. 10362–10374, Dec. 2024

    work page 2024

  4. [4]

    Confidence-driven region mixing for optical remote sensing domain adaptation object detection,

    C. Liu, Y . Dong, Y . Zhang, and X. Li, “Confidence-driven region mixing for optical remote sensing domain adaptation object detection,”IEEE Trans. Geosci. Remote Sensing, vol. 62, pp. 1–14, 2024

    work page 2024

  5. [5]

    Relational part-aware learning for complex composite object detection in high-resolution remote sensing images,

    S. Yuan, L. Zhang, R. Dong, J. Xiong, J. Zheng, H. Fu, and P. Gong, “Relational part-aware learning for complex composite object detection in high-resolution remote sensing images,”IEEE T. Cybern., vol. 54, pp. 6118–6131, Oct. 2024

    work page 2024

  6. [6]

    Learning cross-modality high-resolution representation for thermal small-object detection,

    Y . Zhang, X. Lei, Q. Hu, C. Xu, W. Yang, and G.-S. Xia, “Learning cross-modality high-resolution representation for thermal small-object detection,”IEEE Trans. Geosci. Remote Sensing, vol. 62, pp. 1–15, 2024

    work page 2024

  7. [7]

    Fdgnet: Frequency disentanglement and data geometry for domain generalization in cross- scene hyperspectral image classification,

    B. Qin, S. Feng, C. Zhao, B. Xi, W. Li, and R. Tao, “Fdgnet: Frequency disentanglement and data geometry for domain generalization in cross- scene hyperspectral image classification,”IEEE Trans. Neural Netw. Learn. Syst., pp. 1–14, 2024

    work page 2024

  8. [8]

    Channelwise and spatially guided multimodal feature fusion network for 3-d object detection in autonomous vehicles,

    M. Uzair, J. Dong, R. Shi, H. Mushtaq, and I. Ullah, “Channelwise and spatially guided multimodal feature fusion network for 3-d object detection in autonomous vehicles,”IEEE Trans. Geosci. Remote Sensing, vol. 62, pp. 1–15, 2024

    work page 2024

Show all 66 references
  1. [9]

    Ssmm: Semi-supervised man- ifold method with spatial-spectral self-training and regularized metric constraints for hyperspectral image dimensionality reduction,

    B. Zhu, Y . Jin, X. Guan, and Y . Dong, “Ssmm: Semi-supervised man- ifold method with spatial-spectral self-training and regularized metric constraints for hyperspectral image dimensionality reduction,”Int. J. Appl. Earth Obs. Geoinf., vol. 136, p. 104373, Feb. 2025

    2025

  2. [10]

    Ctnet: Contrastive transformer network for polyp segmentation,

    B. Xiao, J. Hu, W. Li, C.-M. Pun, and X. Bi, “Ctnet: Contrastive transformer network for polyp segmentation,”IEEE T. Cybern., vol. 54, pp. 5040–5053, Sept. 2024

    2024

  3. [11]

    Deep metric learning based on brownian covariance representation for few-shot hyperspectral image classification,

    Y . Dong, B. Zhu, X. Yang, and X. Ma, “Deep metric learning based on brownian covariance representation for few-shot hyperspectral image classification,”IEEE Trans. Geosci. Remote Sensing, vol. 63, pp. 1–13, 2025

    2025

  4. [12]

    Tiny object detection in remote sensing images based on object reconstruction and multiple receptive field adaptive feature enhancement,

    D. Liu, J. Zhang, Y . Qi, Y . Wu, and Y . Zhang, “Tiny object detection in remote sensing images based on object reconstruction and multiple receptive field adaptive feature enhancement,”IEEE Trans. Geosci. Remote Sensing, vol. 62, pp. 1–13, 2024

    2024

  5. [13]

    Multilevel fine- grained features-based general framework for object detection,

    F. Zuo, J. Liu, Z. Chen, H. Zhang, M. Fu, and L. Wang, “Multilevel fine- grained features-based general framework for object detection,”IEEE T. Cybern., vol. 54, pp. 6921–6933, Nov. 2024

    2024

  6. [14]

    Visible and infrared image fusion using deep learning,

    X. Zhang and Y . Demiris, “Visible and infrared image fusion using deep learning,”IEEE Trans. Pattern Anal. Mach. Intell., vol. 45, pp. 10535– 10554, Aug. 2023

    2023

  7. [15]

    Locality guided cross-modal feature aggregation and pixel-level fusion for multispectral pedestrian detection,

    Y . Cao, X. Luo, J. Yang, Y . Cao, and M. Y . Yang, “Locality guided cross-modal feature aggregation and pixel-level fusion for multispectral pedestrian detection,”Inf. Fusion, vol. 88, pp. 1–11, Dec. 2022

    2022

  8. [16]

    Metafusion: Infrared and visible image fusion via meta-feature embedding from object detection,

    W. Zhao, S. Xie, F. Zhao, Y . He, and H. Lu, “Metafusion: Infrared and visible image fusion via meta-feature embedding from object detection,” inProc. CVPR, pp. 13955–13965, June 2023

    2023

  9. [17]

    Rgb-t object detection via group shuffled multi-receptive at- tention and multi-modal supervision,

    J. Wang, X. Tian, S. Dai, T. Zhuo, H. Zeng, H. Liu, J. Liu, X. Zhang, and Y . Zhang, “Rgb-t object detection via group shuffled multi-receptive at- tention and multi-modal supervision,”arXiv preprint arXiv:2405.18955, 2024

    2024 arXiv

  10. [18]

    Damsdet: Dynamic adaptive multispectral detection transformer with competitive query selection and adaptive feature fusion,

    J. Guo, C. Gao, F. Liu, D. Meng, and X. Gao, “Damsdet: Dynamic adaptive multispectral detection transformer with competitive query selection and adaptive feature fusion,” inProc. ECCV, pp. 464–481, Springer, 2025

    2025

  11. [19]

    Translation, scale and rotation: cross- modal alignment meets rgb-infrared vehicle detection,

    M. Yuan, Y . Wang, and X. Wei, “Translation, scale and rotation: cross- modal alignment meets rgb-infrared vehicle detection,” inProc. ECCV, pp. 509–525, Springer, 2022

    2022

  12. [20]

    C 2 former: Calibrated and complementary transformer for rgb-infrared object detection,

    M. Yuan and X. Wei, “C 2 former: Calibrated and complementary transformer for rgb-infrared object detection,”IEEE Trans. Geosci. Remote Sensing, vol. 62, pp. 1–12, 2024

    2024

  13. [21]

    Airborne small target detection method based on multimodal and adaptive feature fusion,

    S. Xu, X. Chen, H. Li, T. Liu, Z. Chen, H. Gao, and Y . Zhang, “Airborne small target detection method based on multimodal and adaptive feature fusion,”IEEE Trans. Geosci. Remote Sensing, vol. 62, pp. 1–15, 2024

    2024

  14. [22]

    Fusion-mamba for cross-modality object detection,

    W. Dong, H. Zhu, S. Lin, X. Luo, Y . Shen, X. Liu, J. Zhang, G. Guo, and B. Zhang, “Fusion-mamba for cross-modality object detection,”arXiv preprint arXiv:2404.09146, 2024

    2024 arXiv

  15. [23]

    Mask-guided mamba fusion for drone-based visible-infrared vehicle detection,

    S. Wang, C. Wang, C. Shi, Y . Liu, and M. Lu, “Mask-guided mamba fusion for drone-based visible-infrared vehicle detection,”IEEE Trans. Geosci. Remote Sensing, vol. 62, pp. 1–12, 2024

    2024

  16. [24]

    Dmm: Disparity-guided multispectral mamba for oriented object detection in remote sensing,

    M. Zhou, T. Li, C. Qiao, D. Xie, G. Wang, N. Ruan, L. Mei, and Y . Yang, “Dmm: Disparity-guided multispectral mamba for oriented object detection in remote sensing,”arXiv preprint arXiv:2407.08132, 2024

    2024 arXiv

  17. [25]

    Drone-based rgb-infrared cross- modality vehicle detection via uncertainty-aware learning,

    Y . Sun, B. Cao, P. Zhu, and Q. Hu, “Drone-based rgb-infrared cross- modality vehicle detection via uncertainty-aware learning,”IEEE Trans. Circuits Syst. Video Technol., vol. 32, pp. 6700–6713, Oct. 2022

    2022

  18. [26]

    Llvip: A visible-infrared paired dataset for low-light vision,

    X. Jia, C. Zhu, M. Li, W. Tang, and W. Zhou, “Llvip: A visible-infrared paired dataset for low-light vision,” inProc. ICCV, pp. 3496–3504, Oct. 2021

    2021

  19. [27]

    Drone- based rgbt tiny person detection,

    Y . Zhang, C. Xu, W. Yang, G. He, H. Yu, L. Yu, and G.-S. Xia, “Drone- based rgbt tiny person detection,”ISPRS-J. Photogramm. Remote Sens., vol. 204, pp. 61–76, 2023

    2023

  20. [28]

    Feature pyramid networks for object detection,

    T.-Y . Lin, P. Doll ´ar, R. Girshick, K. He, B. Hariharan, and S. Be- longie, “Feature pyramid networks for object detection,” inProc. CVPR, pp. 2117–2125, July 2017

    2017

  21. [29]

    Path aggregation network for instance segmentation,

    S. Liu, L. Qi, H. Qin, J. Shi, and J. Jia, “Path aggregation network for instance segmentation,” inProc. CVPR, pp. 8759–8768, June 2018

    2018

  22. [30]

    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 T. Cybern., vol. 52, pp. 8574–8586, Aug. 2022

    2022

  23. [31]

    Face detection with different scales based on faster r-cnn,

    W. Wu, Y . Yin, X. Wang, and D. Xu, “Face detection with different scales based on faster r-cnn,”IEEE T. Cybern., vol. 49, pp. 4017–4028, Nov. 2019

    2019

  24. [32]

    Yolo-world: Real-time open-vocabulary object detection,

    T. Cheng, L. Song, Y . Ge, W. Liu, X. Wang, and Y . Shan, “Yolo-world: Real-time open-vocabulary object detection,” inProc. CVPR, pp. 16901– 16911, June 2024

    2024

  25. [33]

    Object detection under rainy conditions for autonomous vehicles: A review of state-of-the-art and emerging techniques,

    M. Hnewa and H. Radha, “Object detection under rainy conditions for autonomous vehicles: A review of state-of-the-art and emerging techniques,”IEEE Signal Process. Mag., vol. 38, pp. 53–67, Jan. 2021

    2021

  26. [34]

    Unmanned aerial vehicle visual detection and tracking using deep neural networks: A performance benchmark,

    B. K. S. Isaac-Medina, M. Poyser, D. Organisciak, C. G. Willcocks, T. P. Breckon, and H. P. H. Shum, “Unmanned aerial vehicle visual detection and tracking using deep neural networks: A performance benchmark,” inProc. ICCV, pp. 1223–1232, October 2021

    2021

  27. [35]

    Superyolo: Super resolution assisted object detection in multimodal remote sensing imagery,

    J. Zhang, J. Lei, W. Xie, Z. Fang, Y . Li, and Q. Du, “Superyolo: Super resolution assisted object detection in multimodal remote sensing imagery,”IEEE Trans. Geosci. Remote Sensing, vol. 61, pp. 1–15, 2023

    2023

  28. [36]

    Weakly aligned feature fusion for multimodal object detection,

    L. Zhang, Z. Liu, X. Zhu, Z. Song, X. Yang, Z. Lei, and H. Qiao, “Weakly aligned feature fusion for multimodal object detection,”IEEE Trans. Neural Netw. Learn. Syst., pp. 1–15, 2021

    2021

  29. [37]

    Lraf- net: Long-range attention fusion network for visible–infrared object detection,

    H. Fu, S. Wang, P. Duan, C. Xiao, R. Dian, S. Li, and Z. Li, “Lraf- net: Long-range attention fusion network for visible–infrared object detection,”IEEE Trans. Neural Netw. Learn. Syst., vol. 35, pp. 13232– 13245, Oct. 2024

    2024

  30. [38]

    Differential feature awareness network within antagonistic learning for infrared-visible object detection,

    R. Zhang, L. Li, Q. Zhang, J. Zhang, L. Xu, B. Zhang, and B. Wang, “Differential feature awareness network within antagonistic learning for infrared-visible object detection,”IEEE Trans. Circuits Syst. Video Technol., vol. 34, pp. 6735–6748, Aug. 2023

    2023

  31. [39]

    Cdc-yolofusion: Leveraging cross-scale dynamic convolution fusion for visible-infrared object detection,

    Z. Wang, X. Liao, J. Yuan, Y . Yao, and Z. Li, “Cdc-yolofusion: Leveraging cross-scale dynamic convolution fusion for visible-infrared object detection,”IEEE T. Intell. V eh., pp. 1–14, 2024. IEEE TRANSACTIONS ON CYBERNETICS 13

    2024

  32. [40]

    Smfanet: A lightweight self-modulation feature aggregation network for efficient image super- resolution,

    M. Zheng, L. Sun, J. Dong, and J. Pan, “Smfanet: A lightweight self-modulation feature aggregation network for efficient image super- resolution,” inProc. ECCV, pp. 359–375, Springer, 2025

    2025

  33. [41]

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

    A. Dosovitskiy, “An image is worth 16x16 words: Transformers for image recognition at scale,”arXiv preprint arXiv:2010.11929, 2020

    2010 arXiv

  34. [42]

    Seadate: Remedy dual-attention transformer with semantic alignment via contrast learning for multimodal object detection,

    S. Dong, Y . Li, W. Xie, J. Zhang, J. Tian, D. Yang, and J. Lei, “Seadate: Remedy dual-attention transformer with semantic alignment via contrast learning for multimodal object detection,”arXiv preprint arXiv:2410.11358, 2024

    arXiv

  35. [43]

    Improving rgb- infrared object detection with cascade alignment-guided transformer,

    M. Yuan, X. Shi, N. Wang, Y . Wang, and X. Wei, “Improving rgb- infrared object detection with cascade alignment-guided transformer,” Inf. Fusion, vol. 105, p. 102246, May 2024

    2024

  36. [44]

    Caff-dino: Multi-spectral object detection transformers with cross-attention features fusion,

    K. Helvig, B. Abeloos, and P. Trouv ´e-Peloux, “Caff-dino: Multi-spectral object detection transformers with cross-attention features fusion,” in Proc. CVPR, pp. 3037–3046, June 2024

    2024

  37. [45]

    Vision mamba: Efficient visual representation learning with bidirectional state space model,

    L. Zhu, B. Liao, Q. Zhang, X. Wang, W. Liu, and X. Wang, “Vision mamba: Efficient visual representation learning with bidirectional state space model,”arXiv preprint arXiv:2401.09417, 2024

    2024 arXiv

  38. [46]

    Dynamic token augmentation mamba for cross-scene classification of hyperspectral image,

    X. Huang, Y . Zhang, F. Luo, and Y . Dong, “Dynamic token augmentation mamba for cross-scene classification of hyperspectral image,”IEEE Trans. Geosci. Remote Sensing, vol. 62, pp. 1–13, 2024

    2024

  39. [47]

    Como: Cross-mamba interaction and offset-guided fusion for multimodal object detection,

    C. Liu, X. Ma, X. Yang, Y . Zhang, and Y . Dong, “Como: Cross-mamba interaction and offset-guided fusion for multimodal object detection,” arXiv preprint arXiv:2412.18076, 2024

    2024 arXiv

  40. [48]

    Cfmw: Cross-modality fusion mamba for multispectral object detection under adverse weather conditions,

    H. Li, Q. Hu, Y . Yao, K. Yang, and P. Chen, “Cfmw: Cross-modality fusion mamba for multispectral object detection under adverse weather conditions,”arXiv preprint arXiv:2404.16302, 2024

    2024 arXiv

  41. [49]

    Remotedet-mamba: A hybrid mamba-cnn network for multi-modal object detection in remote sensing images,

    K. Ren, X. Wu, L. Xu, and L. Wang, “Remotedet-mamba: A hybrid mamba-cnn network for multi-modal object detection in remote sensing images,”arXiv preprint arXiv:2410.13532, 2024

    2024

  42. [50]

    Weakly misalignment-free adaptive feature alignment for uavs-based multimodal object detection,

    C. Chen, J. Qi, X. Liu, K. Bin, R. Fu, X. Hu, and P. Zhong, “Weakly misalignment-free adaptive feature alignment for uavs-based multimodal object detection,” inProc. CVPR, pp. 26836–26845, June 2024

    2024

  43. [51]

    G2da: Geometry- guided dual-alignment learning for rgb-infrared person re-identification,

    L. Wan, Z. Sun, Q. Jing, Y . Chen, L. Lu, and Z. Li, “G2da: Geometry- guided dual-alignment learning for rgb-infrared person re-identification,” Pattern Recognit., vol. 135, p. 109150, Mar. 2023

    2023

  44. [52]

    Cross-modal oriented object detection of uav aerial images based on image feature,

    H. Wang, C. Wang, Q. Fu, D. Zhang, R. Kou, Y . Yu, and J. Song, “Cross-modal oriented object detection of uav aerial images based on image feature,”IEEE Trans. Geosci. Remote Sensing, vol. 62, pp. 1–21, 2024

    2024

  45. [53]

    Daff: Dual attentive feature fusion for multispectral pedestrian detection,

    A. Althoupety, L.-Y . Wang, W.-C. Feng, and B. Rekabdar, “Daff: Dual attentive feature fusion for multispectral pedestrian detection,” inProc. CVPR, pp. 2997–3006, June 2024

    2024

  46. [54]

    Multimodal object detection by channel switching and spatial attention,

    Y . Cao, J. Bin, J. Hamari, E. Blasch, and Z. Liu, “Multimodal object detection by channel switching and spatial attention,” inProc. CVPR, pp. 403–411, June 2023

    2023

  47. [55]

    Dpnet: Dual- path network for real-time object detection with lightweight attention,

    Q. Zhou, H. Shi, W. Xiang, B. Kang, and L. J. Latecki, “Dpnet: Dual- path network for real-time object detection with lightweight attention,” IEEE Trans. Neural Netw. Learn. Syst., pp. 1–15, 2024

    2024

  48. [56]

    Shufflenet: An extremely efficient convolutional neural network for mobile devices,

    X. Zhang, X. Zhou, M. Lin, and J. Sun, “Shufflenet: An extremely efficient convolutional neural network for mobile devices,” inProc. CVPR, pp. 6848–6856, June 2018

    2018

  49. [57]

    Shufflenet v2: Practi- cal guidelines for efficient cnn architecture design,

    N. Ma, X. Zhang, H.-T. Zheng, and J. Sun, “Shufflenet v2: Practi- cal guidelines for efficient cnn architecture design,” inProc. ECCV, pp. 116–131, 2018

    2018

  50. [58]

    ultralytics/yolov5: v6. 2-yolov5 classification models, apple m1, reproducibility, clearml and deci. ai integrations,

    G. Jocher, A. Chaurasia, A. Stoken, J. Borovec, Y . Kwon, K. Michael, J. Fang, C. Wong, Z. Yifu, D. Montes,et al., “ultralytics/yolov5: v6. 2-yolov5 classification models, apple m1, reproducibility, clearml and deci. ai integrations,”Zenodo, 2022

    2022

  51. [59]

    Visible-thermal uav tracking: A large-scale benchmark and new baseline,

    P. Zhang, J. Zhao, D. Wang, H. Lu, and X. Ruan, “Visible-thermal uav tracking: A large-scale benchmark and new baseline,” inProc. CVPR, pp. 8886–8895, June 2022

    2022

  52. [60]

    Cross-modality fu- sion transformer for multispectral object detection,

    F. Qingyun, H. Dapeng, and W. Zhaokui, “Cross-modality fu- sion transformer for multispectral object detection,”arXiv preprint arXiv:2111.00273, 2021

    2021 arXiv

  53. [61]

    Guided hybrid quantization for object detection in remote sensing imagery via one-to- one self-teaching,

    J. Zhang, J. Lei, W. Xie, Y . Li, G. Yang, and X. Jia, “Guided hybrid quantization for object detection in remote sensing imagery via one-to- one self-teaching,”IEEE Trans. Geosci. Remote Sensing, vol. 61, pp. 1– 15, 2023

    2023

  54. [62]

    Misaligned rgb-infrared object detection via adaptive dual-discrepancy calibration,

    M. He, Q. Wu, K. N. Ngan, F. Jiang, F. Meng, and L. Xu, “Misaligned rgb-infrared object detection via adaptive dual-discrepancy calibration,” Remote Sens., vol. 15, no. 19, p. 4887, 2023

    2023

  55. [63]

    Icafusion: Iterative cross-attention guided feature fusion for multispectral object detection,

    J. Shen, Y . Chen, Y . Liu, X. Zuo, H. Fan, and W. Yang, “Icafusion: Iterative cross-attention guided feature fusion for multispectral object detection,”Pattern Recognit., vol. 145, p. 109913, Jan. 2024

    2024

  56. [64]

    Gm-detr: Generalized muiltispectral detection transformer with efficient fusion encoder for visible-infrared detection,

    Y . Xiao, F. Meng, Q. Wu, L. Xu, M. He, and H. Li, “Gm-detr: Generalized muiltispectral detection transformer with efficient fusion encoder for visible-infrared detection,” inProc. CVPR, pp. 5541–5549, June 2024

    2024

  57. [65]

    Unveiling the limits of alignment: Multi-modal dynamic local fusion network and a benchmark for unaligned rgbt video object detection,

    Q. Wang, Z. Tu, K. Wang, L. Gu, and C. Guo, “Unveiling the limits of alignment: Multi-modal dynamic local fusion network and a benchmark for unaligned rgbt video object detection,”arXiv preprint arXiv:2410.12143, 2024

    2024 arXiv

  58. [66]

    Aodet: Anti-occlusion for enhanced small object detection in drone-based rgbt imagery,

    Z. Gui, Y . Zhang, X. Lei, R. Zhang, and W. Yang, “Aodet: Anti-occlusion for enhanced small object detection in drone-based rgbt imagery,” in Proc. IGARSS, pp. 954–958, July 2024

    2024

Pith tools

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