skip to main content
10.1145/3633637.3633679acmotherconferencesArticle/Chapter ViewAbstractPublication PagesiccprConference Proceedingsconference-collections
research-article

Meta-ResNet: A Novel Few-shot SAR Target Recognition Method Based on Meta-learning

Published: 28 February 2024 Publication History

Abstract

Automatic target recognition (ATR) based on synthetic aperture radar (SAR) has drawn extensive attention. Deep neural networks based data-driven recognition algorithms have been reported as feasible and promising methods for SAR ATR in recent years. However, most deep models usually need a large number of annotated data for parameter optimization, otherwise they would possibly encounter a serious overfitting problem when limited training examples per class are available. To solve mentioned problems, a novel few-shot SAR target recognition method named Meta-ResNet is proposed, which successfully combines residual networks (ResNet) with meta-learning framework. In the proposed method, we design a novel learner based on ResNet, which is more suitable for processing sparse SAR images because of the residual mapping structure. Furthermore, the improved meta-learner in this paper can learn a good initialization for learner as well as proper but different learning rates for every parameter of learner with different meta learning rates. Consequently, the trained improved meta-learner can guide the learner to converge well and quickly, which is significant to improve the generalization performance with only few labeled training samples. Experiments are conducted on the Moving and Stationary Target Acquisition and Recognition (MSTAR) dataset to evaluate the proposed method. Compared with other three few-shot learning SAR target recognition methods, the experimental results showed the superiority of our method.

References

[1]
Baiyuan Ding and Gongjian Wen. 2018. Target Reconstruction Based on 3-D Scattering Center Model for Robust SAR ATR. IEEE Transactions on Geoscience and Remote Sensing 56, 7 (2018), 3772–3785. https://doi.org/10.1109/TGRS.2018.2810181
[2]
Chelsea Finn, Pieter Abbeel, and Sergey Levine. 2017. Model-agnostic meta-learning for fast adaptation of deep networks. In Proc. Int. Conf. Mach. Learn. (ICML). 1199–1208.
[3]
Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. 2016. Deep Residual Learning for Image Recognition. In 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). 770–778. https://doi.org/10.1109/CVPR.2016.90
[4]
Z. Li, F. Zhou, C. Fei, and L. Hang. 2017. Meta-SGD: Learning to Learn Quickly for Few-Shot Learning. (2017).
[5]
Gerard Margarit, Jordi J. Mallorqui, Joaquim Fortuny-Guasch, and Carlos Lopez-Martinez. 2009. Exploitation of Ship Scattering in Polarimetric SAR for an Improved Classification Under High Clutter Conditions. IEEE Transactions on Geoscience and Remote Sensing 47, 4 (2009), 1224–1235. https://doi.org/10.1109/TGRS.2008.2008721
[6]
F. Qin, J. Guo, and W. Sun. 2017. Object-oriented ensemble classification for polarimetric SAR Imagery using restricted Boltzmann machines. Remote Sensing Letters 8, 1-3 (2017), 204–213.
[7]
Asma Raeisi, Gholamreza Akbarizadeh, and Alimorad Mahmoudi. 2018. Combined Method of an Efficient Cuckoo Search Algorithm and Nonnegative Matrix Factorization of Different Zernike Moment Features for Discrimination Between Oil Spills and Lookalikes in SAR Images. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing 11, 11 (2018), 4193–4205. https://doi.org/10.1109/JSTARS.2018.2841503
[8]
Sachin Ravi and Hugo Larochelle. 2017. Optimization as a model for few-shot learning. In Proc. Int. Conf. Learn. Represent. (ICLR).
[9]
Umamahesh Srinivas, Vishal Monga, and Raghu G. Raj. 2014. SAR Automatic Target Recognition Using Discriminative Graphical Models. IEEE Trans. Aerospace Electron. Systems 50, 1 (2014), 591–606. https://doi.org/10.1109/TAES.2013.120340
[10]
Chanika Sukawattanavijit, Jie Chen, and Hongsheng Zhang. 2017. GA-SVM Algorithm for Improving Land-Cover Classification Using SAR and Optical Remote Sensing Data. IEEE Geoscience and Remote Sensing Letters 14, 3 (2017), 284–288. https://doi.org/10.1109/LGRS.2016.2628406
[11]
Yijun Sun, Zhipeng Liu, Sinisa Todorovic, and Jian Li. 2007. Adaptive boosting for SAR automatic target recognition. IEEE Trans. Aerospace Electron. Systems 43, 1 (2007), 112–125. https://doi.org/10.1109/TAES.2007.357120
[12]
Oriol Vinyals, Charles Blundell, Timothy Lillicrap, Daan Wierstra, 2016. Matching networks for one shot learning. In Advances in neural information processing systems. 3630–3638.
[13]
Jianlong Wang, Biao Hou, Licheng Jiao, and Shuang Wang. 2020. POL-SAR Image Classification Based on Modified Stacked Autoencoder Network and Data Distribution. IEEE Transactions on Geoscience and Remote Sensing 58, 3 (2020), 1678–1695. https://doi.org/10.1109/TGRS.2019.2947633
[14]
Jinsong Zhang, Mengdao Xing, and Yiyuan Xie. 2021. FEC: A Feature Fusion Framework for SAR Target Recognition Based on Electromagnetic Scattering Features and Deep CNN Features. IEEE Transactions on Geoscience and Remote Sensing 59, 3 (2021), 2174–2187. https://doi.org/10.1109/TGRS.2020.3003264
[15]
X. Zhang, J. Zhang, J. I. Yong-Gang, and J. M. Meng. 2010. The capability analysis of ship classification by structure feature using SAR images. Acta Oceanologica Sinica (2010).
[16]
Z. Zhao, L. Jiao, J. Zhao, J. Gu, and Z. Jin. 2016. Discriminant Deep Belief Network for High-Resolution SAR Image Classification. Pattern Recognition 61 (2016), 686–701.

Index Terms

  1. Meta-ResNet: A Novel Few-shot SAR Target Recognition Method Based on Meta-learning
        Index terms have been assigned to the content through auto-classification.

        Recommendations

        Comments

        Information & Contributors

        Information

        Published In

        cover image ACM Other conferences
        ICCPR '23: Proceedings of the 2023 12th International Conference on Computing and Pattern Recognition
        October 2023
        589 pages
        ISBN:9798400707988
        DOI:10.1145/3633637
        Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. Copyrights for components of this work owned by others than the author(s) must be honored. Abstracting with credit is permitted. To copy otherwise, or republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee. Request permissions from [email protected].

        Publisher

        Association for Computing Machinery

        New York, NY, United States

        Publication History

        Published: 28 February 2024

        Permissions

        Request permissions for this article.

        Check for updates

        Author Tags

        1. automatic target recognition (ATR)
        2. few-shot learnin
        3. meta-learning
        4. residual networks
        5. synthetic aperture radar (SAR)

        Qualifiers

        • Research-article
        • Research
        • Refereed limited

        Funding Sources

        Conference

        ICCPR 2023

        Contributors

        Other Metrics

        Bibliometrics & Citations

        Bibliometrics

        Article Metrics

        • 0
          Total Citations
        • 33
          Total Downloads
        • Downloads (Last 12 months)33
        • Downloads (Last 6 weeks)5
        Reflects downloads up to 20 Feb 2025

        Other Metrics

        Citations

        View Options

        Login options

        View options

        PDF

        View or Download as a PDF file.

        PDF

        eReader

        View online with eReader.

        eReader

        HTML Format

        View this article in HTML Format.

        HTML Format

        Figures

        Tables

        Media

        Share

        Share

        Share this Publication link

        Share on social media