Journal of Korean Society of Industrial and Systems Engineering (산업경영시스템학회지)

Society of Korea Industrial and System Engineering (한국산업경영시스템학회)
권호 표지
DOI QR코드

DOI QR Code

Underwater Sonar Image Classification based on Vision Transformer with Metric Learning

비전 트랜스포머 및 메트릭 러닝 기반 수중 소나 이미지 분류

  • Received : 2025.11.19
  • Accepted : 2025.12.16
  • Published : 2025.12.31
Download PDF
⟨ Previous Next ⟩

Abstract

Underwater sonar image classification is essential for maritime surveillance, autonomous navigation, and underwater target identification, where optical sensing is often restricted by turbidity and light attenuation. To enhance the robustness of sonar-based perception under such challenging conditions, this study proposes a metric-enhanced Vision Transformer (ViT) framework that integrates Siamese-based representation alignment with distance-regularized classification. In the first stage, a Siamese pre-training strategy is employed to align embeddings of positive pairs, encouraging directionally consistent representations that improve class separability even under severe noise and viewpoint variations. In the second stage, the pretrained ViT encoder is frozen, and five classifiers-Linear, Cosine, Proxy, and their Mahalanobis-regularized variants-are systematically evaluated to investigate the effect of embedding normalization and distributional alignment. Experimental results on the UATD dataset demonstrate that the Siamese-trained ViT produces more stable and discriminative features than both ResNet-50 and standard ViT-S. Among the classifiers, the Mahalanobis-regularized cosine classifier achieves the highest, showing significant reductions in misclassification between visually similar classes such as cube and square cage. Overall, the proposed approach highlights the effectiveness of combining ViT with metric learning and covariance-aware distance normalization for underwater sonar image recognition. The results suggest that metric-enhanced transformers offer a robust and generalizable foundation for sonar-based perception in real maritime environments.

Keywords

Acknowledgement

This work was supported by the KRIT (Korea Research Institute for Defense Technology Planning and Advancement) and LIG Nex1, grant funded by the Defense Acquisition Program Administration (DAPA).

References

  1. Caron, M., Touvron, H., Misra, I., Jégou, H., Mairal, J., Bojanowski, P., and Joulin, A., Emerging Properties in Self-Supervised Vision Transformers, Proceedings of the IEEE/CVF International Conference on Computer Vision, 2021, pp. 9650-9660.
  2. Chen, X. and He, K., Exploring Simple Siamese Representation Learning, Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2021, pp. 15750-15758.
  3. Choi, H., Lee, J., and Kim, M., Automatic Mine Detection in Side-Scan Sonar Images Using Machine Learning Approaches, IEEE Journal of Oceanic Engineering, 2018, Vol. 43, No. 4, pp. 1020-1035.
  4. Chungath, T.T., Nambiar, A.M., and Mittal, A., Transfer Learning and Few-Shot Learning Based Deep Neural Network Models for Underwater Sonar Image Classification With a Few Samples, IEEE Journal of Oceanic Engineering, 2024, Vol 49, No. 1, pp. 294-310. https://doi.org/10.1109/JOE.2022.3221127
  5. Domingos, L.C., Santos, P.E., Skelton, P.S., Brinkworth, R.S., and Sammut, K., A survey of underwater acoustic data classification methods using deep learning for shoreline surveillance, Sensors, 2022, Vol. 22, No. 6, Article: 2181. https://doi.org/10.3390/s22062181
  6. Dosovitskiy, A., Beyer, L., Kolesnikov, A., Weissenborn, D., Zhai, X., Unterthiner, T., Dehghani, M., Minderer, M., Heigold, G., Gelly, S. and Uszkoreit, J., An image is worth 16x16 words: Transformers for image recognition at scale, arXiv:2010.11929, 2020.
  7. Fan, Z., Xia, W., Liu, X. and Li, H., Detection and segmentation of underwater objects from forward-looking sonar based on a modified mask RCNN, Signal, Image Video Process, 2021, Vol. 15, pp. 1135-1143. https://doi.org/10.1007/s11760-020-01841-x
  8. Fuchs, L.R., Gallstrom, A., and Folkesson, J., Object recognition in forward looking sonar images using transfer learning, 2018 IEEE/OES Autonomous Underwater Vehicle Workshop (AUV), 2018, pp. 1-6.
  9. Gong, M., Chen, C., Sun, Q., Wang, Y., and Huang, H., Out-of-distribution detection with prototypical outlier proxy, Proceedings of the AAAI Conference on Artificial Intelligence, 2025, Vol. 39, No. 16, pp. 16835-16843.
  10. Jin, L., Liang, H. and Yang, C., Accurate underwater ATR in forward-looking sonar imagery using deep convolutional neural networks, IEEE Access, Vol. 7, 2019, pp. 125522-125531. https://doi.org/10.1109/Access.6287639
  11. Kim, S., Kim, D., Cho, M., and Kwak, S., Proxy anchor loss for deep metric learning, Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2020, pp. 3238-3247.
  12. Koch, G., Zemel, R., and Salakhutdinov, R., Siamese neural networks for one-shot image recognition, ICML Deep Learning Workshop, 2015, Vol. 2, No. 1, pp. 1-30.
  13. Movshovitz-Attias, Y., Toshev, A., Leung, T.K., Ioffe, S., and Singh, S., No fuss distance metric learning using proxies, Proceedings of the IEEE International Conference on Computer Vision, 2017, pp. 360-368.
  14. Rajani, H., Gracias, N., and Garcia, R., A Convolutional Vision Transformer for Semantic Segmentation of Side-Scan Sonar Data, arXiv:2302.12416, 2023.
  15. Regmi, S., Panthi, B., Ming, Y., Gyawali, P.K., Stoyanov, D., and Bhattarai, B., Reweightood: Loss reweighting for distance-based ood detection, Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2024, pp. 131-141.
  16. Schroff, F., Kalenichenko, D., and Philbin, J, Facenet: A unified embedding for face recognition and clustering, Proceedings of the IEEE Conference on Computer, 2015, pp. 815-823.
  17. Steiniger, Y., Kraus, D., and Meisen, T., Survey on deep learning based computer vision for sonar imagery, Engineering Applications of Artificial Intelligence, 2022, Vol. 114, Article 105157. https://doi.org/10.1016/j.engappai.2022.105157
  18. Sun, Y., Zheng, H., Zhang, G., Ren, J., Xu, H., and Xu, C., DP-ViT: A Dual-Path Vision Transformer for Real-Time Sonar Target Detection, Remote Sensing, 2022, Vol. 14, No. 22, Article: 5807. https://doi.org/10.3390/rs14225807
  19. Valdenegro-Toro, M., Object recognition in forward-looking sonar images with convolutional neural networks, OCEANS 2016 MTS/IEEE Monterey, 2016, pp. 1-6.
  20. Vasankari, L., Borzyszkowski, A., Zelioli, L., and Heikkonen, J., Deep Mix: AI in Littoral Sonar Operations, 2025, Journal of Marine Science and Application, 1-12.
  21. Xie, K., Yang, J., and Qiu, K., A Dataset with Multibeam Forward-Looking Sonar for Underwater Object Detection, arXiv:2212.00352, 2022.
  22. Yang, J. and Xie, K. Underwater acoustic target detection (UATD) dataset. https://doi.org/10.6084/m9.figshare.21331143.v3 , 2022.
  23. Zhu, X., Liang, Y., Zhang, J., and Chen, Z., STAFNet : Swin transformer based anchor-free network for detection of forward-looking sonar imagery, Proceedings of the 2022 International Conference on Multimedia Retrieval, 2022, pp. 443-450.