volume | PIER Journals

Abstract

Synthetic Aperture Radar (SAR) Automatic Target Recognition (ATR) often suffers from significant performance degradation under Extended Operating Conditions (EOCs), including depression-angle shifts, target configuration variations, and complex background clutter. Most convolutional neural network (CNN)-based methods represent SAR images using first-order Euclidean features, which may be insufficient to explicitly model structured interactions among compact feature dimensions. To address this limitation, we propose a structured SPD-based shared-branch SAR ATR framework. Specifically, a shared CNN backbone processes three label-preserving branch inputs to obtain branch feature maps. After feature projection, global average pooling, and L2 normalization, a regularized outer-product matrix is constructed for each branch to explicitly encode pairwise multiplicative feature interactions. The resulting positive-definite matrix is compactly represented by direct upper-triangular vectorization. Descriptors from the three branches are subsequently integrated via globally shared Softmax scaling and concatenation before classification. The evaluation reports five independent runs and compares vector, covariance, Log-SPD, and parameter-matched nonlinear representations. The controlled analyses reveal protocol-dependent behavior, with the outer-product representation showing its clearest relative advantage under depression-angle variation. For settings using repeated branch inputs, no complementary-view claim is made. Overall, the results support the structured outer-product representation within the evaluated MSTAR protocols.

1. Keydel, Eric R., Shung W. Lee, and John T. Moore, "MSTAR extended operating conditions: A tutorial," Algorithms for Synthetic Aperture Radar Imagery III, Vol. 2757, 228-242, Orlando, Florida, USA, June 1996.
doi:10.1117/12.242059

2. Ross, Timothy D., Steven W. Worrell, Vincent J. Velten, John C. Mossing, and Michael L. Bryant, "Standard SAR ATR evaluation experiments using the MSTAR public release data set," Algorithms for Synthetic Aperture Radar Imagery V, Vol. 3370, 566-573, Orlando, Florida, USA, September 1998.
doi:10.1117/12.321859

3. El-Darymli, Khalid, Eric W. Gill, Peter Mcguire, Desmond Power, and Cecilia Moloney, "Automatic target recognition in synthetic aperture radar imagery: A state-of-the-art review," IEEE Access, Vol. 4, 6014-6058, 2016.
doi:10.1109/access.2016.2611492 Google Scholar

4. Kechagias-Stamatis, Odysseas and Nabil Aouf, "Automatic target recognition on synthetic aperture radar imagery: A survey," IEEE Aerospace and Electronic Systems Magazine, Vol. 36, No. 3, 56-81, March 2021.
doi:10.1109/maes.2021.3049857 Google Scholar

5. Slesinski, Jakub and Damian Wierzbicki, "Review of synthetic aperture radar automatic target recognition: A dual perspective on classical and deep learning techniques," IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, Vol. 18, 18978-19024, 2025.
doi:10.1109/jstars.2025.3589804 Google Scholar

6. Yin, Junjun, Changxian Duan, Hongbo Wang, and Jian Yang, "A review on the few-shot SAR target recognition," IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, Vol. 17, 16411-16425, 2024.
doi:10.1109/jstars.2024.3454266 Google Scholar

7. Yang, Ying and Haitao Zhao, "Deep learning-based SAR target recognition: A dual-perspective survey of closed set and open set," Applied Sciences, Vol. 15, No. 23, 12501, November 2025.
doi:10.3390/app152312501 Google Scholar

8. Wen, Zaidao, Youlan Yu, and Qian Wu, "Multimodal discriminative feature learning for SAR ATR: A fusion framework of phase history, scattering topology, and image," IEEE Transactions on Geoscience and Remote Sensing, Vol. 62, 1-14, 2024.
doi:10.1109/tgrs.2023.3340651 Google Scholar

9. Lv, Benyuan, Ying Luo, Jiacheng Ni, Siyuan Zhao, Jia Liang, Yingxi Liu, and Qun Zhang, "MMFF: Multiview and multi-level feature fusion method within limited sample conditions for SAR image target recognition," ISPRS Journal of Photogrammetry and Remote Sensing, Vol. 224, 302-316, June 2025.
doi:10.1016/j.isprsjprs.2025.03.010 Google Scholar

10. Choi, Jae-Ho, Myung-Jun Lee, Nam-Hoon Jeong, Geon Lee, and Kyung-Tae Kim, "Fusion of target and shadow regions for improved SAR ATR," IEEE Transactions on Geoscience and Remote Sensing, Vol. 60, 1-17, 2022.
doi:10.1109/tgrs.2022.3165849 Google Scholar

11. Xiao, Ziyi, Gong Zhang, and Qijun Dai, "Multiview features centers sample expansion for SAR image classification," IEEE Geoscience and Remote Sensing Letters, Vol. 21, 1-5, 2024.
doi:10.1109/lgrs.2024.3350668 Google Scholar

12. He, Xinlin, Chao Li, Kaiming Li, and Ying Luo, "SDF-net: A space-frequency dynamic fusion network for SARATR," Progress In Electromagnetics Research B, Vol. 115, 25-37, 2025.
doi:10.2528/pierb25060702 Google Scholar

13. Tong, Xiaobao and Yong Wang, "Active multiview fusion framework for SAR automatic target recognition," IEEE Transactions on Instrumentation and Measurement, Vol. 73, 1-14, 2024.
doi:10.1109/tim.2023.3341140 Google Scholar

14. Zhang, Tingwei, "A Multi-view SAR target recognition method based on adaptive weighted decision fusion," Remote Sensing Letters, Vol. 14, No. 11, 1196-1205, November 2023.
doi:10.1080/2150704x.2023.2277157 Google Scholar

15. Hu, Zhe, Gong Zhang, Yong Liu, and Hongchao Liu, "Cross-view SAR target recognition based on multi-view correlation and multi-task architecture," Remote Sensing Letters, Vol. 14, No. 7, 743-753, July 2023.
doi:10.1080/2150704x.2023.2235637 Google Scholar

16. Wang, Zijun, Gong Zhang, Daiyin Zhu, and Qijun Dai, "MV-ResFPN: A residual network deeply combining multi-view and multiscale fusion for SAR target recognition," Sixteenth International Conference on Graphics and Image Processing (ICGIP 2024), Vol. 13539, 135390, February 2025.
doi:10.1117/12.3057774

17. Novak, L. M., G. J. Owirka, and W. S. Brower, "Performance of 10-and 20-target MSE classifiers," IEEE Transactions on Aerospace and Electronic Systems, Vol. 36, No. 4, 1279-1289, October 2000.
doi:10.1109/7.892675 Google Scholar

18. Song, Shengli, Bin Xu, and Jian Yang, "SAR target recognition via supervised discriminative dictionary learning and sparse representation of the SAR-HOG feature," Remote Sensing, Vol. 8, No. 8, 683, August 2016.
doi:10.3390/rs8080683 Google Scholar

19. Chen, Sizhe, Haipeng Wang, Feng Xu, and Ya-Qiu Jin, "Target classification using the deep convolutional networks for SAR images," IEEE Transactions on Geoscience and Remote Sensing, Vol. 54, No. 8, 4806-4817, August 2016.
doi:10.1109/tgrs.2016.2551720 Google Scholar

20. Simonyan, K. and A. Zisserman, "Very deep convolutional networks for large-scale image recognition," arXiv preprint arXiv:1409.1556, 2014.
doi:10.48550/arXiv.1409.1556 Google Scholar

21. Szegedy, Christian, W. Liu, Y. Jia, et al., "Going deeper with convolutions," 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 1-9, Boston, MA, USA, June 2015.
doi:10.1109/cvpr.2015.7298594

22. He, Kaiming, Xiangyu Zhang, Shaoqing Ren, and Jian Sun, "Deep residual learning for image recognition," 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 770-778, June 2016.
doi:10.1109/cvpr.2016.90

23. Li, Chao, Jiacheng Ni, Ying Luo, Dan Wang, and Qun Zhang, "A dual-branch spatial-frequency domain fusion method with cross attention for SAR image target recognition," Remote Sensing, Vol. 17, No. 14, 2378, July 2025.
doi:10.3390/rs17142378 Google Scholar

24. Tuzel, Oncel, Fatih Porikli, and Peter Meer, "Region covariance: A fast descriptor for detection and classification," Computer Vision --- ECCV 2006, 589-600, Springer, Berlin, Heidelberg, 2006.
doi:10.1007/11744047_45

25. Pennec, Xavier, Pierre Fillard, and Nicholas Ayache, "A riemannian framework for tensor computing," International Journal of Computer Vision, Vol. 66, No. 1, 41-66, January 2006.
doi:10.1007/s11263-005-3222-z Google Scholar

26. Huang, Zhiwu and Luc Van Gool, "A riemannian network for SPD matrix learning," Proceedings of the AAAI Conference on Artificial Intelligence, 2036-2042, February 2017.
doi:10.1609/aaai.v31i1.10866

27. Wang, R., X.-J. Wu, H. Li, and J. Kittler, "Riemannian self-attention mechanism for SPD networks," arXiv preprint arXiv:2311.16738, 2023. Google Scholar

28. Chen, Ziheng, Tianyang Xu, Xiao-Jun Wu, Rui Wang, Zhiwu Huang, and Josef Kittler, "Riemannian local mechanism for SPD neural networks," Proceedings of the AAAI Conference on Artificial Intelligence, Vol. 37, No. 6, 7104-7112, June 2023.
doi:10.1609/aaai.v37i6.25867

29. Wang, Rui, Xiao-Jun Wu, Tianyang Xu, Cong Hu, and Josef Kittler, "U-SPDNet: An SPD manifold learning-based neural network for visual classification," Neural Networks, Vol. 161, 382-396, April 2023.
doi:10.1016/j.neunet.2022.11.030 Google Scholar

30. Wang, Rui, Xiao-Jun Wu, Ziheng Chen, Tianyang Xu, and Josef Kittler, "Dreamnet: A deep riemannian manifold network for SPD matrix learning," Computer Vision --- ACCV 2022, 646-663, Springer, Cham, 2023.
doi:10.1007/978-3-031-26351-4_39

31. Arsigny, Vincent, Pierre Fillard, Xavier Pennec, and Nicholas Ayache, "Log-Euclidean metrics for fast and simple calculus on diffusion tensors," Magnetic Resonance in Medicine, Vol. 56, No. 2, 411-421, June 2006.
doi:10.1002/mrm.20965 Google Scholar

32. Arsigny, Vincent, Pierre Fillard, Xavier Pennec, and Nicholas Ayache, "Geometric means in a novel vector space structure on symmetric positive‐definite matrices," SIAM Journal on Matrix Analysis and Applications, Vol. 29, No. 1, 328-347, January 2007.
doi:10.1137/050637996 Google Scholar

33. Szegedy, Christian, Vincent Vanhoucke, Sergey Ioffe, Jon Shlens, and Zbigniew Wojna, "Rethinking the inception architecture for computer vision," 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2818-2826, Las Vegas, NV, USA, June 2016.
doi:10.1109/cvpr.2016.308

34. Loshchilov, I. and F. Hutter, "Decoupled weight decay regularization," 7th International Conference on Learning Representations, ICLR 2019, New Orleans, LA, USA, May 2019.

35. Loshchilov, I. and F. Hutter, "SGDR: Stochastic gradient descent with warm restarts," arXiv preprint arXiv:1608.03983, April 2017.
doi:10.48550/arXiv.1608.03983

36. Kingma, D. P. and J. Ba, "Adam: A method for stochastic optimization," arXiv preprint arXiv:1412.6980, 2015.
doi:10.48550/arXiv.1412.6980

37. Smith, L. N., "A disciplined approach to neural network hyper-parameters: Part 1 --- Learning rate, batch size, momentum, and weight decay," arXiv preprint arXiv:1803.09820, 2018. Google Scholar

38. Shorten, Connor and Taghi M. Khoshgoftaar, "A survey on image data augmentation for deep learning," Journal of Big Data, Vol. 6, 60, July 2019. Google Scholar

39. Paszke, A., S. Gross, F. Massa, A. Lerer, J. Bradbury, G. Chanan, T. Killeen, Z. Lin, N. Gimelshein, L. Antiga, A. Desmaison, A. Köpf, E. Yang, Z. DeVito, M. Raison, A. Tejani, S. Chilamkurthy, B. Steiner, L. Fang, J. Bai, and S. Chintala, "Pytorch: An imperative style, high-performance deep learning library," 33rd Conference on Neural Information Processing Systems (NeurIPS 2019), Vancouver, Canada, 2019.

40. Zhang, Linbin, Xiangguang Leng, Sijia Feng, Xiaojie Ma, Kefeng Ji, Gangyao Kuang, and Li Liu, "Domain knowledge powered two-stream deep network for few-shot SAR vehicle recognition," IEEE Transactions on Geoscience and Remote Sensing, Vol. 60, 1-15, 2022.
doi:10.1109/tgrs.2021.3116349 Google Scholar

41. Wang, Chenwei, Yulin Huang, Xiaoyu Liu, Jifang Pei, Yin Zhang, and Jianyu Yang, "Global in local: A convolutional transformer for SAR ATR FSL," IEEE Geoscience and Remote Sensing Letters, Vol. 19, 1-5, 2022.
doi:10.1109/lgrs.2022.3183467 Google Scholar

42. Wang, Chenwei, Siyi Luo, Yulin Huang, Jifang Pei, Yin Zhang, and Jianyu Yang, "SAR ATR method with limited training data via an embedded feature augmenter and dynamic hierarchical-feature refiner," IEEE Transactions on Geoscience and Remote Sensing, Vol. 61, 1-15, 2023.
doi:10.1109/tgrs.2023.3314501 Google Scholar

43. Maaten, L. and G. Hinton, "Visualizing data using t-SNE," Journal of Machine Learning Research, Vol. 9, No. 86, 2579-2605, 2008. Google Scholar

Ms. Tong Xie

Dr. Jiacheng Ni

Dr. Jia Liang

Dr. Chao Li

Dr. Yubin Qin