Sonar image recognition plays a crucial role in the field of underwater environment detection. While existing sonar image recognition models based on deep neural network have improved classification accuracy, they often face the challenges of long-tailed distribution in practice, leading to insufficient identification of rare yet high-value targets. To remedy this, we propose a novel Multi-expert Dynamic Collaboration model to enhance recognition accuracy for long-tailed sonar image (MEDC-SI). Our model consists of multi-expert network and dynamic learning strategy. The multi-expert network contains a conventional branch for feature representation learning and two re-balancing branches for tail samples learning. And three experts collaborate to achieve imbalanced sonar image recognition. The dynamic learning strategy is designed to shift the focus of model training between the conventional branch and re-balancing branches to improve the feature learning and classifier learning simultaneously. Finally, extensive experimental results on three sonar image recognition datasets, KLSG, FLSMDD, and NKSID, demonstrate the effectiveness of the proposed model, achieving overall accuracies of 91.51%, 99.74%, and 96.19%, respectively.
HUANGH N, LIB Q, LIUJ Y, et al. Sonar Image Underwater Target Recognition: A Comprehensive Overview and Prospects[J]. J Electron Inf Technol, 2024, 46(5): 1742-1760. DOI: 10.11999/JEIT231207 .
VALDENEGRO-TOROM, PRECIADO-GRIJALVAA, WEHBEB. Pre-trained Models for Sonar Images[C]//OCEANS 2021. San Diego-Porto: IEEE, 2021: 1-8. DOI: 10.23919/OCEANS44145.2021.9705825 .
[5]
CHOH, GUJ, YUS C. Robust Sonar-based Underwater Object Recognition Against Angle-of-view Variation[J]. IEEE Sens J, 2016, 16(4): 1013-1025. DOI: 10.1109/JSEN.2015.2496945 .
[6]
MARSZALJ, SALAMONR. Detection Range of Intercept Sonar for CWFM Signals[J]. Arch Acoust, 2015, 39(2): 215-230. DOI: 10.2478/aoa-2014-0026 .
[7]
FEIT, KRAUSD, ZOUBIRA M. Contributions to Automatic Target Recognition Systems for Underwater Mine Classification[J]. IEEE Trans Geosci Remote Sens, 2015, 53(1): 505-518. DOI: 10.1109/TGRS.2014.2324971 .
[8]
MYERSV, FAWCETTJ. A Template Matching Procedure for Automatic Target Recognition in Synthetic Aperture Sonar Imagery[J]. IEEE Signal Process Lett, 2010, 17(7): 683-686. DOI: 10.1109/LSP.2010.2051574 .
[9]
BARNGROVERC M. Automated Detection of Mine-like Objects in Side Scan Sonar Imagery[M]. San Diego: University of California, 2014.
[10]
CHHABRAP S, PAILHASY, WALLACEA M, et al. Target Classification in SAS Imagery Using Orthogonal Basis Selection[C]//OCEANS 2017-Aberdeen. New York: IEEE, 2017: 1-10. DOI: 10.1109/OCEANSE.2017.8084647 .
[11]
WILLIAMSD P. Underwater Target Classification in Synthetic Aperture Sonar Imagery Using Deep Convolutional Neural Networks[C]//2016 23rd International Conference on Pattern Recognition (ICPR). New York: IEEE, 2016: 2497-2502. DOI: 10.1109/ICPR.2016.7900011 .
[12]
CHENGZ, HUOG Y, LIH S. A Multi-domain Collaborative Transfer Learning Method with Multi-scale Repeated Attention Mechanism for Underwater Side-scan Sonar Image Classification[J]. Remote Sens, 2022, 14(2): 355. DOI: 10.3390/rs14020355 .
[13]
XUH P, YANGL Z, ZHANGM X. Unsupervised Classification Based on Deep Adaptation Network for Sonar Images[J]. J Electron Imag, 2023, 32: 013029. DOI: 10.1117/1.JEI.32.1.013029 .
[14]
JIAOW P, ZHANGJ L. Sonar Images Classification while Facing Long-tail and Few-shot[J]. IEEE Trans Geosci Remote Sens, 2022, 60: 4210420. DOI: 10.1109/TGRS.2022.3211847 .
YANGY X, WEIY, GUOJ, et al. Recognition of Side-scan Sonar Images under Long-tail Distribution[C]//International Conference on Computer Vision and Image Processing (CVIP 2024). Bellingham: SPIE, 2025: 33. DOI: 10.1117/12.3058318 .
[17]
LIS B, MAJ F, WUY L, et al. SSS Small Target Detection via Combining Weighted Sparse Model with Shadow Characteristics[J]. IEEE Trans Geosci Remote Sens, 2023, 61: 5911711. DOI: 10.1109/TGRS.2023.3285436 .
[18]
ZHUP P, ISAACSJ, FUB, et al. Deep Learning Feature Extraction for Target Recognition and Classification in Underwater Sonar Images[C]//2017 IEEE 56th Annual Conference on Decision and Control (CDC). New York: IEEE, 2017: 2724-2731. DOI: 10.1109/CDC.2017.8264055 .
[19]
HEK M, ZHANGX Y, RENS Q, et al. Deep Residual Learning for Image Recognition[C]//2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). New York: IEEE, 2016: 770-778. DOI: 10.1109/CVPR.2016.90 .
[20]
HOWARDA G, ZHUM, CHENB, et al. Mobilenets: Efficient Convolutional Neural Networks for Mobile Vision Applications[EB/OL]. (2017-04-17)[2025-03-15].
[21]
HUANGG, LIUZ, VAN DER MAATENL, et al. Densely Connected Convolutional Networks[C]//2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). New York: IEEE, 2017: 2261-2269. DOI: 10.1109/CVPR.2017.243 .
[22]
ARRIAGAO, VALDENEGRO-TOROM, PLÖGERP. Real-time Convolutional Neural Networks for Emotion and Gender Classification[J]. (2017-10-20)[2025-03-15].
[23]
YANGZ W, ZHAOJ H, YUY C, et al. A Sample Augmentation Method for Side-scan Sonar Full-class Images that Can be Used for Detection and Segmentation[J]. IEEE Trans Geosci Remote Sens, 2024, 62: 5908111. DOI: 10.1109/TGRS.2024.3371051 .
[24]
XUH L, BAIZ Y, ZHANGX Y, et al. MFSANet: Zero-shot Side-scan Sonar Image Recognition Based on Style Transfer[J]. IEEE Geosci Remote Sens Lett, 2023, 20: 1503105. DOI: 10.1109/LGRS.2023.3318051 .
CHENY L, LIB, LIANGH, et al. Research on Sonar Image Few-shot Classification Based on Deep Learning[J]. J Northwest Polytech Univ, 2022, 40(4): 739-745. DOI: org/10.1051/jnwpu/20224040739 .
[27]
ZHOUH, LUOT J, HEY M. Dynamic Collaborative Learning with Heterogeneous Knowledge Transfer for Long-tailed Visual Recognition[J]. Inf Fusion, 2025, 115: 102734. DOI: 10.1016/j.inffus.2024.102734 .
[28]
DUC Q, WANGY L, SONGS J, et al. Probabilistic Contrastive Learning for Long-tailed Visual Recognition[J]. IEEE Trans Pattern Anal Mach Intell, 2024, 46(9): 5890-5904. DOI: 10.1109/TPAMI.2024.3369102 .
[29]
ZHOUB Y, CUIQ, WEIX S, et al. BBN: Bilateral-branch Network with Cumulative Learning for Long-tailed Visual Recognition[C]//2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). New York: IEEE, 2020: 9716-9725. DOI: 10.1109/cvpr42600.2020.00974 .
[30]
RENJ, YUC, MAX, et al. Balanced Meta-softmax for Long-tailed Visual Recognition[C]//Advances in Neural Information Processing Systems. Red Hook: Curran Associates. 2020, 33: 4175-4186.
[31]
GAOJ, ZHAOH, LIZ, et al. Enhancing Minority Classes by Mixing: An Adaptative Optimal Transport Approach for Long-tailed Classification[C]//Advances in Neural Information Processing Systems. Red Hook: Curran Associates. 2023, 36: 60329-60348.
[32]
LIM K, HUZ K, LUY, et al. Feature Fusion from Head to Tail for Long-tailed Visual Recognition[J]. Proc AAAI Conf Artif Intell, 2024, 38(12): 13581-13589. DOI: 10.1609/aaai.v38i12.29262 .
[33]
ZHANGH, CISSEM, DAUPHINY N, et al. Mixup: Beyond Empirical Risk Minimization[C/OL]//International Conference on Learning Representations, 2018. (2017-11-06)[2025-03-15].
[34]
CHOUH P, CHANGS C, PANJ Y, et al. Remix: Rebalanced Mixup[M]//Computer Vision-ECCV 2020 Workshops. Cham: Springer International Publishing, 2020: 95-110. DOI: 10.1007/978-3-030-65414-6_9 .
[35]
TANJ R, WANGC B, LIB Y, et al. Equalization Loss for Long-tailed Object Recognition[C]//2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). New York: IEEE, 2020: 11659-11668. DOI: 10.1109/cvpr42600.2020.01168 .
[36]
JAMALM A, BROWNM, YANGM H, et al. Rethinking Class-balanced Methods for Long-tailed Visual Recognition from a Domain Adaptation Perspective[C]//2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). New York: IEEE, 2020: 7607-7616. DOI: 10.1109/CVPR42600.2020.00763 .
[37]
HUANGC, ZHAOJ H, ZHANGH M, et al. Seg2Sonar: a Full-class Sample Synthesis Method Applied to Underwater Sonar Image Target Detection, Recognition, and Segmentation Tasks[J]. IEEE Trans Geosci Remote Sens, 2024, 62: 5909319. DOI: 10.1109/TGRS.2024.3363875 .
[38]
HUOG Y, WUZ Y, LIJ B. Underwater Object Classification in Sidescan Sonar Images Using Deep Transfer Learning and Semisynthetic Training Data[J]. IEEE Access, 2020, 8: 47407-47418. DOI: 10.1109/ACCESS.2020.2978880 .
[39]
VALDENEGRO-TOROM. Deep Neural Networks for Marine Debris Detection in Sonar Images[EB/OL]. (2019-05-13) [2025-03-15].
[40]
YEX F, LIC L, ZHANGS Y, et al. Research on Side-scan Sonar Image Target Classification Method Based on Transfer Learning[C]//OCEANS 2018 MTS/IEEE Charleston. New York: IEEE, 2018: 1-6. DOI: 10.1109/OCEANS.2018.8604691 .
[41]
WANGY X, RAMANAND, HEBERTM. Learning to Model the Tail[C]//Advances in Neural Information Processing Systems. Red Hook: Curran Associates. 2017: 30.
[42]
CAOK, WEIC, GAIDONA, et al. Learning Imbalanced Datasets with Label-distribution-aware Margin Loss[C]//Advances in Neural Information Processing Systems. Red Hook: Curran Associates. 2019: 32.
[43]
ALSHAMMARIS, WANGY X, RAMANAND, et al. Long-Tailed Recognition via Weight Balancing[C]//2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). New York: IEEE, 2022: 6887-6897. DOI: 10.1109/CVPR52688.2022.00677 .