Pith. sign in

REVIEW 3 major objections 6 minor 41 references

DRPFNet: Dual-domain Residual Progressive Fusion Network for RGB-Thermal Object Detection

T0 review · 3 major / 6 minor · reviewed 2026-08-15 · deepseek-v4-flash

Pith's one-line read DRPFNet claims that RGB-thermal detection improves when fusion is progressive, bidirectional, and split across frequency and spatial domains.

desk verdict A cleanly ablated incremental fusion architecture whose headline SOTA claim is built on unrerun cross-protocol numbers; worth refereeing, not worth accepting at face value. read the letter →

arxiv 2608.03370 v1 pith:TCI6BPGM submitted 2026-08-04 cs.CV

classification cs.CV
keywords RGB-thermalobjectdetectiondual-domainfusionfrequency-domainfeatureseparationprogressivecross-scaleedge-guidedattentionbidirectionalenhancementLLVIPdatasetM3FD
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper sets out to show that RGB-thermal object detection is better served by a progressive, bidirectional, dual-domain fusion architecture than by the independent scale-wise attention or transformer merges used in most current detectors. DRPFNet organizes this around three linked modules: MRF-BFE propagates downsampled fused features from lower to higher scales and lets fused and modality-specific features refine each other; DFAF separates RGB and thermal spectra into high- and low-frequency bands with edge guidance; EGMK applies edge-guided strip convolutions at the finest detection level. On LLVIP the network reports 97.8 mAP50 and 64.4 mAP, and on M3FD 88.7 mAP50 and 61.8 mAP, ahead of the published numbers compared in the paper. The authors attribute the gains to treating RGB and thermal as spectrally complementary, with RGB edges and thermal structure rather than two spatial images to be mixed.

What carries the argument

The load-bearing object is the DFAF module's frequency-domain band separation: a learnable radius partitions the 2D FFT spectrum into low- and high-frequency components, and the inverse-FFT bands are blended with Scharr edge features through a learned channel gate, with a residual connection. This operationalizes the paper's premise that RGB carries high-frequency edge and texture while thermal carries low-frequency structure. Around it, MRF-BFE creates the progressive fusion spine by concatenating each level's RGB and thermal features with the downsampled fused output of the previous level, then adding bidirectional refinement at P3/P4, and EGMK applies the same dual-domain idea at the detection stage by using edge-extracted attention maps to guide strip convolutions and dilated convolutions. Together the three modules form a single information-flow system spanning extraction, fusion, and detection.

What would settle it

Run the top compared detectors, such as ICAFusion, Fusion-Mamba, and MMFN, in the same codebase as DRPFNet with the same YOLO11 backbones, 640 by 640 inputs, epochs, and augmentations, and compare mAP50 and mAP on LLVIP and M3FD; if the gaps shrink to within run-to-run variance, the paper's central claim is refuted.

Watch

Extended reading notes

Core claim

The central claim is that three structural weaknesses of existing RGB-T detectors—fusion done independently at each feature scale, a one-way fusion-to-detection flow, and spatial-only feature mixing—can be fixed jointly by one network, and that fixing them improves detection accuracy on standard benchmarks. DRPFNet's MRF-BFE makes fusion progressive: each level receives the downsampled fused output of the previous level, so geometric detail constrains semantic fusion, and bidirectional pathways at P3/P4 let fused and modality-specific features refine each other. The DFAF module splits each modality's features into low- and high-frequency bands using a learnable radius in the 2D FFT domain, extracts Scharr edge features in the spatial domain, and blends both through a channel-wise gate with a residual connection. The EGMK module, placed at P3, uses edge-guided attention to steer horizontal and vertical strip convolutions and cascaded dilated convolutions toward object boundaries, sharpening foreground-background separation. The paper's experiments and ablations attribute each reported improvement—the jump from baseline 84.8 to 88.7 mAP50 on M3FD, and from 94.2 to 97.8 on LLVIP—to the progressive addition of these modules.

Load-bearing premise

The strongest claim depends on comparing DRPFNet's numbers with each prior method's self-reported numbers, which were produced under different backbones, input resolutions, and training schedules; if those differences account for the reported 1 to 3 point gaps, the claimed state-of-the-art superiority would not survive a controlled re-run.

Editorial extensions

If this is right

  • If DRPFNet's numbers hold under matched protocols, it shows that explicit frequency-domain band separation plus edge guidance can beat attention-only and transformer-only fusion on RGB-T detection.
  • Progressive residual fusion predicts that localization at stricter IoU thresholds improves most, because low-level geometric priors constrain high-level semantic fusion; the reported mAP gains on M3FD are consistent with that.
  • Thermal low-frequency structure should keep detections stable in low light, so the method is expected to degrade less than spatial-only fusion when visible images become dark or noisy, as on LLVIP.
  • EGMK's edge-guided strip convolutions should specifically help with occluded, partially visible, or thin objects, a concrete checkable prediction for future analysis.
  • The complexity analysis says the modules keep the same order of computational cost as the baseline detector, so the accuracy gains are not bought by a qualitatively heavier architecture.

Reading between the lines

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

  • A natural transfer is to other paired modalities with the same spectral asymmetry, such as RGB-depth or RGB-event, where one sensor supplies texture and the other supplies structure; the DFAF gating logic could carry over without redesign.
  • The paper uses a single learnable radius for band separation; an extension would make the radius scale-dependent or input-dependent, which would test whether the benefit comes from the specific frequency split or from the gating and residual machinery.
  • The ablation results suggest thermal-only bidirectional enhancement helps more than RGB-only; an input-dependent illumination estimator could make the enhancement weights adaptive per image pair.
  • Applying the same progressive fusion spine to video RGB-T detection, with temporal frames as an additional domain, is a direct extension the paper does not explore.
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

3 major / 6 minor

Summary. The paper introduces DRPFNet, a dual-stream RGB-thermal detection network built on YOLO11, with three designed modules: MRF-BFE (multi-level residual fusion with bidirectional feature enhancement), DFAF (dual-domain adaptive frequency/spatial fusion in the backbone), and EGMK (edge-guided multi-scale kernel placed at the P3 detection level). The method is evaluated on LLVIP and M3FD, reporting mAP50/mAP values of 97.8/64.4 and 88.7/61.8, respectively. The experimental section includes a four-step ablation on both datasets, an efficiency comparison, and Grad-CAM visualizations. The central claim is that the three modules jointly improve detection accuracy, making the network competitive with or superior to recent RGB-T detectors.

Significance. If the reported gains hold under fair comparison, the work is of interest because it explicitly targets three underexplored issues in RGB-T detection: cross-scale knowledge propagation, bidirectional optimization, and frequency-spatial collaboration. The ablation study in Table IV is internally consistent and shows that each module contributes a monotonic improvement on both datasets, which is a useful empirical signal for the community. The complexity analysis in Section III.E is a strength. The significance is conditional, however, because the state-of-the-art comparison in Tables I and II mixes training protocols, and the absolute margins over some competitors are small.

major comments (3)
  1. [Section IV.B, Tables I and II] The external comparison is not controlled: all starred numbers are copied from the original papers, which use different backbones (Faster R-CNN, YOLOv5, etc.), different input resolutions, and different training schedules. For instance, on LLVIP DRPFNet's mAP50 of 97.8 is only 0.3 points above the 97.5 reported for CFT, and on M3FD the mAP of 61.8 is 2.5 points above MRD-YOLO's 59.3. These gaps are within the range that protocol differences can produce. The claim that DRPFNet outperforms state-of-the-art methods is therefore not established. The authors should either re-run the compared methods under their own protocol or explicitly soften the claim to 'competitive with published numbers' and discuss the protocol limitation.
  2. [Section IV.D, Table IV] No measure of variance is reported: each ablation configuration is a single run, and some adjacent differences are as small as 0.8 mAP50 (85.6 vs 84.8 on M3FD). Without multiple seeds or statistical testing, the module contributions cannot be distinguished from run-to-run noise. The main ablation table should include at least three seeds with standard deviations, or a statistical significance test, to support the central claim that each module is beneficial.
  3. [Section III.B, Eq. (1)] The learnable band separation in DFAF is not fully specified. The text says a learnable radius parameter separates low- and high-frequency components of the FFT spectrum, but it does not explain how the radius is represented, how it is applied to the Fourier coefficients, how it is constrained to a valid range, or how it relates to the weights W[0] and W[1] in Eq. (1). Without this detail, the DFAF module is underspecified and not reproducible. Please provide the exact mathematical formulation, including any normalization or thresholding.
minor comments (6)
  1. [Abstract vs. Section IV.B] The abstract states 'competitive performance with competitive efficiency' while Section IV.B repeatedly says 'outperforming' recent methods; please align the wording to avoid contradiction.
  2. [Figure 1] Figure 1 is difficult to read because the interconnections between the three modules and the backbone are visually tangled; please redraw it at higher magnification with clearer labeling of the pathways.
  3. [Section III.B, Eq. (2)] Eq. (2) uses a residual term X without defining it; state explicitly that X is the input feature map to the DFAF block.
  4. [Section IV.A] The implementation details state that gradient accumulation is used, but the number of accumulation steps is not given; please specify this value.
  5. [Section IV.E] The explanation for the scooter/rider case in the Grad-CAM analysis is presented as fact; it is a hypothesis and should be labeled as such.
  6. [References] References [4] and [26] are the same paper (Cross-modality fusion transformer); please deduplicate.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: DRPFNet's results come from direct experiments and ablations, not from fitting inputs renamed as predictions or from load-bearing self-citations.

full rationale

The paper's central claim is an empirical performance comparison on LLVIP and M3FD. Its proposed modules (MRF-BFE, DFAF, EGMK) are architectural components defined by explicit equations, and the reported mAP values are measured experimental outcomes rather than quantities derived from fitted parameters or from the definitions of those modules. The ablation study in Table IV compares variants of the model against a clearly specified baseline, and each incremental gain is an observed result, not a construction-level identity. Tables I and II copy prior results from their original papers, as indicated by the asterisk notation; this creates a legitimate concern about protocol comparability across different backbones, resolutions, and training schedules, but that is a benchmarking-validity issue, not a circularity issue, because DRPFNet's own numbers are independently produced under its stated protocol. No self-citation is load-bearing for the main contribution: the references to related work, including the cross-modality fusion transformer by authors with overlapping names, are contextual and do not supply the proof of DRPFNet's effectiveness. There is no uniqueness theorem, no ansatz smuggled in by citation, and no renaming of a known empirical pattern as a new result. The qualitative explanations in Sections IV.B and IV.E are post-hoc narrative about why the modules help; they are not derivations that reduce to their own assumptions. Overall, the paper is an empirical systems paper with self-contained ablations, and no circular step is identifiable from the text.

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

The central empirical claim depends on standard CV structures (YOLO11, C3k2, FPN-PAN) taken from sources, a domain assumption about RGB-thermal frequency complementarity, and a handful of hand-chosen hyperparameters; no physically invented entities are introduced. The learnable radius in DFAF is a fitted parameter whose converged value is not reported, and the standard hyperparameters are not swept, which limits the strength of the efficiency claims.

free parameters (4)
  • Learnable band separation radius in DFAF = not reported
    Chosen by network training; no analysis of the converged value or sensitivity to it is provided.
  • Input resolution = 640x640
    Standard choice, but no comparison against other resolutions is given.
  • EGMK dilation rates = 2 and 4
    Hand-selected; no ablation of alternate dilation settings is reported.
  • Training epochs = 300 for M3FD, 100 for LLVIP
    Standard choices; no sweep or early stopping analysis is reported.
assumptions (4)
  • domain assumption YOLO11 is a strong baseline detector and its features are suitable for RGB-T fusion.
    Used as the dual-stream backbone; no comparison with other backbones is provided.
  • domain assumption RGB and thermal modalities are complementary in the frequency domain, with RGB holding high-frequency edges and thermal low-frequency contours.
    Adopted from cited prior work [1], [12], [13] and used as the design premise for DFAF; not re-derived or measured in this paper.
  • domain assumption The building blocks C3k2, FPN-PAN, Scharr operator, and strip convolution behave as described in their source papers.
    These are taken as given from YOLO11 [23], FPN [24], and Strip R-CNN [25] references.
  • domain assumption The train/test split of M3FD follows Liang et al. [34] and is compatible with the splits used by prior methods.
    No verification of split compatibility across papers is provided.

how reviews work

0 comments
Cite this review

Pith. "Pith review of DRPFNet: Dual-domain Residual Progressive Fusion Network for RGB-Thermal Object Detection." pith.science (2026). https://pith.science/paper/TCI6BPGM

@misc{pith2026260803370,
  author       = {Pith},
  title        = {Pith review of: DRPFNet: Dual-domain Residual Progressive Fusion Network for RGB-Thermal Object Detection},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/TCI6BPGM}},
  note         = {Machine review of arXiv:2608.03370}
}
read the original abstract

RGB-thermal (RGB-T) object detection aims to fuse complementary information from visible and thermal modalities to achieve robust detection under varying illumination and weather conditions. Current methods typically employ attention mechanisms or transformers to perform cross-modal fusion independently at each feature scale, directly combining RGB and thermal features in the spatial domain. However, they still face significant limitations: cross-level knowledge inheritance caused by independent fusion at each scale,suppressing noise continuously due to the lack of bidirectional optimization, and information degradation induced by the absence of frequency-spatial collaboration. To address these issues, we propose DRPFNet, a Dual-domain Residual Progressive Fusion Network that constructs a unified information flow optimization system from three synergistic levels:structure, feature, and enhancement. At the structural level, we establish cross-scale propagation through bottom-up knowledge accumulation and bidirectional enhancement,ensuring smooth information flow. At the feature level, we collaboratively extract RGB high-frequency edges and thermal low-frequency structures via frequency band separation and edge guidance, guaranteeing representation quality. At the enhancement level, we enhance foreground-background discrimination through edge-guided dual-domain refinement,achieving precise object localization.Extensive experiments on two public RGB-T datasets demonstrate that our method achieves competitive performance with competitive efficiency, validating the effectiveness of this hierarchical collaborative strategy.

Figures

Figures reproduced from arXiv: 2608.03370 by the authors.

Figure 1
Figure 1. Overview of the proposed DRPFNet for RGB-thermal object detection. The network consists of three main components: [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Grad-CAM visualization comparing baseline and full [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

41 extracted references · 36 canonical work pages

  1. [1]

    Infrared and visible image fusion methods and applica- tions: A survey,

    Y . Maet al., “Infrared and visible image fusion methods and applica- tions: A survey,”Inf. Fusion, vol. 45, pp. 153–178, 2019

  2. [2]

    Cross- modality interactive attention network for multispectral pedestrian de- tection,

    L. Zhang, Z. Liu, S. Zhang, X. Yang, H. Qiao, K. Huanget al., “Cross- modality interactive attention network for multispectral pedestrian de- tection,”Inf. Fusion, vol. 50, pp. 20–29, 2019

  3. [3]

    Improving multispectral pedestrian detection by addressing modality imbalance problems,

    K. Zhou, L. Chen, and X. Cao, “Improving multispectral pedestrian detection by addressing modality imbalance problems,” inECCV, 2020, pp. 787–803

  4. [5]

    Illumination-aware faster R- CNN for robust multispectral pedestrian detection,

    C. Li, D. Song, R. Tong, and M. Tang, “Illumination-aware faster R- CNN for robust multispectral pedestrian detection,”Pattern Recognit., vol. 85, pp. 161–171, 2019

  5. [6]

    HAFNet: Hierarchical attentive fusion network for multispectral pedestrian detection,

    P. Peng, T. Xu, B. Huang, and J. Li, “HAFNet: Hierarchical attentive fusion network for multispectral pedestrian detection,”Remote Sens., vol. 15, no. 8, p. 2041, 2023

  6. [7]

    Target-aware dual adversarial learning and a multi-scenario multi-modality benchmark to fuse infrared and visible for object detection,

    J. Liu, X. Fan, Z. Huang, G. Wu, R. Liu, W. Zhonget al., “Target-aware dual adversarial learning and a multi-scenario multi-modality benchmark to fuse infrared and visible for object detection,” inCVPR, 2022, pp. 5792–5801

  7. [8]

    High performance RGB-thermal video object detection via hybrid fusion with progressive interaction and temporal-modal difference,

    Q. Wang, Z. Tu, C. Li, and J. Tang, “High performance RGB-thermal video object detection via hybrid fusion with progressive interaction and temporal-modal difference,”Inf. Fusion, vol. 114, p. 102665, 2025

  8. [9]

    Guided attentive feature fusion for multispectral pedestrian detection,

    H. Zhang, E. Fromont, S. Lef `evre, and B. Avignon, “Guided attentive feature fusion for multispectral pedestrian detection,” inWACV, 2021, pp. 72–80

Show all 41 references
  1. [10]

    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, 2024

  2. [11]

    Multidimensional fusion network for multispectral object detection,

    F. Yang, B. Liang, W. Li, and J. Zhang, “Multidimensional fusion network for multispectral object detection,”IEEE Trans. Circuits Syst. Video Technol., vol. 35, no. 1, pp. 547–560, 2025

  3. [12]

    MCFusion: Frequency domain characteristics enhancement and feature compensation fusion network for RGB-T object detection,

    Y . Gao, Z. Liao, Y . Liu, A. Yi, and G. Zhang, “MCFusion: Frequency domain characteristics enhancement and feature compensation fusion network for RGB-T object detection,”IEEE Sens. J., vol. 25, no. 11, pp. 20 880–20 893, 2025

  4. [13]

    Spectral-aware global fusion for RGB-thermal semantic segmentation,

    C. Zhang, Z. Wan, S. Stepputtis, K. Sycara, and Y . Xie, “Spectral-aware global fusion for RGB-thermal semantic segmentation,”arXiv preprint arXiv:2505.15491, 2025

  5. [14]

    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,” inCVPRW, 2023, pp. 403–411

  6. [15]

    Multi-scale aggregation transformers for multispectral object detection,

    S. You, X. Xie, Y . Feng, C. Mei, and Y . Ji, “Multi-scale aggregation transformers for multispectral object detection,”IEEE Signal Process. Lett., vol. 30, pp. 1172–1176, 2023

  7. [16]

    CDDFuse: Correlation-driven dual-branch feature decomposition for multi-modality image fusion,

    Z. Zhao, H. Bai, J. Zhang, Y . Zhang, S. Xu, Z. Linet al., “CDDFuse: Correlation-driven dual-branch feature decomposition for multi-modality image fusion,” inCVPR, 2023, pp. 5906–5916

  8. [17]

    CrossFormer: Cross-guided attention for multi-modal object detection,

    S. Lee, J. Park, and J. Park, “CrossFormer: Cross-guided attention for multi-modal object detection,”Pattern Recognit. Lett., vol. 179, pp. 144– 150, 2024

  9. [18]

    Fusion- Mamba for cross-modality object detection,

    W. Dong, H. Zhu, S. Lin, X. Luo, Y . Shen, G. Guoet al., “Fusion- Mamba for cross-modality object detection,”IEEE Trans. Multimedia, vol. 27, pp. 7392–7406, 2025

  10. [19]

    WaveMamba: Wavelet-driven mamba fusion for RGB-infrared object detection,

    H. Zhu, W. Dong, L. Yang, H. Li, Y . Yang, Y . Renet al., “WaveMamba: Wavelet-driven mamba fusion for RGB-infrared object detection,” in ICCV, 2025, pp. 11 219–11 229

  11. [20]

    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, 2024

  12. [21]

    Removal then selection: A coarse-to-fine fusion perspective for RGB-infrared object detection,

    T. Zhao, M. Yuan, F. Jiang, N. Wang, and X. Wei, “Removal then selection: A coarse-to-fine fusion perspective for RGB-infrared object detection,”arXiv preprint arXiv:2401.10731, 2024

  13. [22]

    Optimal operators in digital image processing,

    H. Scharr, “Optimal operators in digital image processing,” Ph.D. dissertation, University of Heidelberg, 2000

  14. [23]

    YOLO11 by ultralytics,

    G. Jocher, A. Chaurasia, and J. Qiu, “YOLO11 by ultralytics,” https: //github.com/ultralytics/ultralytics, 2024

  15. [24]

    Feature pyramid networks for object detection,

    T.-Y . Lin, P. Doll´ar, R. Girshick, K. He, B. Hariharan, and S. Belongie, “Feature pyramid networks for object detection,” inCVPR, 2017, pp. 936–944

  16. [25]

    Strip R-CNN: Large strip convolution for remote sensing object detection,

    X. Yuan, Z. Zheng, Y . Li, X. Liu, L. Liu, X. Li, Q. Hou, and M.-M. Cheng, “Strip R-CNN: Large strip convolution for remote sensing object detection,”arXiv preprint arXiv:2501.03775, 2025

  17. [26]

    Cross-modality fusion transformer for multispectral object detection,

    Q. Fang, D. Han, and Z. Wang, “Cross-modality fusion transformer for multispectral object detection,” 2022. [Online]. Available: https: //arxiv.org/abs/2111.00273

  18. [27]

    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,” inCVPR, 2023, pp. 13 955–13 965

  19. [28]

    Diff-IF: Multi-modality image fusion via diffusion model with fusion knowledge prior,

    X. Yi, L. Tang, H. Zhang, H. Xu, and J. Ma, “Diff-IF: Multi-modality image fusion via diffusion model with fusion knowledge prior,”Inf. Fusion, p. 102450, 2024

  20. [29]

    CAMF: An interpretable infrared and visible image fusion network based on class activation mapping,

    L. Tang, Z. Chen, J. Huang, and J. Ma, “CAMF: An interpretable infrared and visible image fusion network based on class activation mapping,”IEEE Trans. Multimedia, vol. 26, pp. 4776–4791, 2024

  21. [30]

    YOLO-Adaptor: A fast adaptive one-stage detector for non-aligned visible-infrared object detection,

    H. Fu, H. Liu, J. Yuan, X. He, J. Lin, and Z. Li, “YOLO-Adaptor: A fast adaptive one-stage detector for non-aligned visible-infrared object detection,”IEEE Trans. Intell. Veh., vol. 9, no. 11, pp. 7070–7083, 2024

  22. [31]

    ACDF-YOLO: Attentive and cross-differential fusion network for multimodal remote sensing object detection,

    X. Fei, M. Guo, Y . Li, R. Yu, and L. Sun, “ACDF-YOLO: Attentive and cross-differential fusion network for multimodal remote sensing object detection,”Remote Sens., vol. 16, no. 18, p. 3532, 2024

  23. [32]

    A dual-modality pedestrian detection method based on multi-scale feature fusion,

    H. Yang, W. Bian, J. Wang, Y . Bo, and Y . Mi, “A dual-modality pedestrian detection method based on multi-scale feature fusion,” in ICHMS, 2024, pp. 1–6

  24. [33]

    TFDet: Target-aware fusion for RGB-T pedestrian detection,

    X. Zhang, X. Zhang, J. Wang, J. Ying, Z. Sheng, H. Yuet al., “TFDet: Target-aware fusion for RGB-T pedestrian detection,”IEEE Trans. Neural Netw. Learn. Syst., vol. 36, no. 7, pp. 13 276–13 290, 2025

  25. [34]

    Explicit attention-enhanced fusion for RGB-thermal perception tasks,

    M. Liang, J. Hu, C. Bao, H. Feng, F. Deng, and T. L. Lam, “Explicit attention-enhanced fusion for RGB-thermal perception tasks,”IEEE Robot. Autom. Lett., vol. 8, pp. 4060–4067, 2023

  26. [35]

    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,” inICCV, 2021, pp. 3496–3504

  27. [36]

    SuperFusion: A ver- satile image registration and fusion network with semantic awareness,

    L. Tang, Y . Deng, Y . Ma, J. Huang, and J. Ma, “SuperFusion: A ver- satile image registration and fusion network with semantic awareness,” IEEE/CAA J. Autom. Sin., vol. 9, no. 12, pp. 2121–2137, 2022

  28. [37]

    Learning a graph neural network with cross modality interaction for image fusion,

    J. Li, J. Chen, J. Liu, and H. Ma, “Learning a graph neural network with cross modality interaction for image fusion,” inACM MM, 2023, pp. 4471–4479

  29. [38]

    KCDNet: Multimodal object detection in modal information imbalance scenes,

    H. Wang, S. Qu, Z. Qiao, and X. Liu, “KCDNet: Multimodal object detection in modal information imbalance scenes,”IEEE Trans. Instrum. Meas., vol. 73, pp. 1–13, 2024

  30. [39]

    MRD-YOLO: A multi- spectral object detection algorithm for complex road scenes,

    C. Sun, Y . Chen, X. Qiu, R. Li, and L. You, “MRD-YOLO: A multi- spectral object detection algorithm for complex road scenes,”Sensors, vol. 24, no. 10, p. 3222, 2024

  31. [40]

    Cross-modal oriented object detection of UA V 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 UA V aerial images based on image feature,”IEEE Trans. Geosci. Remote Sens., vol. 62, pp. 1–21, 2024

  32. [41]

    Equivariant multi-modality image fusion,

    Z. Zhao, H. Bai, J. Zhang, Y . Zhang, K. Zhang, S. Xu, D. Chen, R. Timofte, and L. V . Gool, “Equivariant multi-modality image fusion,” inCVPR, 2024, pp. 25 912–25 921

  33. [42]

    RGB-IR YOLO combining modality-specific reconstruction and information integration,

    Y . Chen, B. Wang, W. Zhu, and J. Yuan, “RGB-IR YOLO combining modality-specific reconstruction and information integration,” inYAC, 2024, pp. 2045–2050

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.