Unsupervised Domain-Adaptive SAR Ship Detection Based on Cross-Domain Feature Interaction and Data Contribution Balance

被引:3
|
作者
Yang, Yanrui [1 ,2 ]
Chen, Jie [1 ,2 ]
Sun, Long [1 ]
Zhou, Zheng [3 ]
Huang, Zhixiang [2 ]
Wu, Bocai [1 ]
机构
[1] China Elect Technol Grp Corp, Res Inst 38, Hefei 230088, Peoples R China
[2] Anhui Univ, Sch Elect & Informat Engn, Key Lab Intelligent Comp & Signal Proc, Minist Educ, Hefei 230601, Peoples R China
[3] Natl Univ Def Technol NUDT, Coll Elect Sci & Technol, Key Lab Complex Electromagnet Environm Effects Ele, Changsha 410073, Peoples R China
基金
中国国家自然科学基金;
关键词
deep learning; SAR; object detection; unsupervised domain adaptation; OBJECT DETECTION; ADAPTATION; ALGORITHM; ALIGNMENT; IMAGES;
D O I
10.3390/rs16020420
中图分类号
X [环境科学、安全科学];
学科分类号
08 ; 0830 ;
摘要
Due to the complex imaging mechanism of SAR images and the lack of multi-angle and multi-parameter real scene SAR target data, the generalization performance of existing deep-learning-based synthetic aperture radar (SAR) image target detection methods are extremely limited. In this paper, we propose an unsupervised domain-adaptive SAR ship detection method based on cross-domain feature interaction and data contribution balance. First, we designed a new cross-domain image generation module called CycleGAN-SCA to narrow the gap between the source domain and the target domain. Second, to alleviate the influence of complex backgrounds on ship detection, a new backbone using a self-attention mechanism to tap the potential of feature representation was designed. Furthermore, aiming at the problems of low resolution, few features and easy information loss of small ships, a new lightweight feature fusion and feature enhancement neck was designed. Finally, to balance the influence of different quality samples on the model, a simple and efficient E12IoU Loss was constructed. Experimental results based on a self-built large-scale optical-SAR cross-domain target detection dataset show that compared with existing cross-domain methods, our method achieved optimal performance, with the mAP reaching 68.54%. Furthermore, our method achieved a 6.27% improvement compared to the baseline, even with only 5% of the target domain labeled data.
引用
收藏
页数:21
相关论文
共 50 条
  • [31] Cross-domain object detection using unsupervised image translation
    Arruda, Vinicius F.
    Berriel, Rodrigo F.
    Paixao, Thiago M.
    Badue, Claudine
    De Souza, Alberto F.
    Sebe, Nicu
    Oliveira-Santos, Thiago
    EXPERT SYSTEMS WITH APPLICATIONS, 2022, 192
  • [32] Cross-domain few-shot learning based on feature adaptive distillation
    Zhang, Dingwei
    Yan, Hui
    Chen, Yadang
    Li, Dichao
    Hao, Chuanyan
    NEURAL COMPUTING & APPLICATIONS, 2024, 36 (08): : 4451 - 4465
  • [33] Domain Generalization and Feature Fusion for Cross-domain Imperceptible Adversarial Attack Detection
    Li, Yi
    Angelov, Plamen
    Suri, Neeraj
    2023 INTERNATIONAL JOINT CONFERENCE ON NEURAL NETWORKS, IJCNN, 2023,
  • [34] Cross-domain few-shot learning based on feature adaptive distillation
    Dingwei Zhang
    Hui Yan
    Yadang Chen
    Dichao Li
    Chuanyan Hao
    Neural Computing and Applications, 2024, 36 : 4451 - 4465
  • [35] Unsupervised domain adaptation segmentation algorithm with cross-domain data augmentation and category contrast
    Dong, Wenyong
    Liang, Zhixue
    Wang, Liping
    Tian, Gang
    Long, Qianhui
    NEUROCOMPUTING, 2025, 623
  • [36] Cross-Domain Label-Adaptive Stance Detection
    Hardalov, Momchil
    Arora, Arnav
    Nakov, Preslav
    Augenstein, Isabelle
    2021 CONFERENCE ON EMPIRICAL METHODS IN NATURAL LANGUAGE PROCESSING (EMNLP 2021), 2021, : 9011 - 9028
  • [37] Cross-Domain Object Detection by Dual Adaptive Branch
    Liu, Xinyi
    Zhang, Baofeng
    Liu, Na
    SENSORS, 2023, 23 (03)
  • [38] Unsupervised Cross-domain Learning by Interaction Information Co-clustering
    Ando, Shin
    Suzuki, Einoshin
    ICDM 2008: EIGHTH IEEE INTERNATIONAL CONFERENCE ON DATA MINING, PROCEEDINGS, 2008, : 13 - +
  • [39] Unsupervised Domain Adaptation Based on Progressive Transfer for Ship Detection: From Optical to SAR Images
    Shi, Yu
    Du, Lan
    Guo, Yuchen
    Du, Yuang
    IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, 2022, 60
  • [40] Striking a Balance: Unsupervised Cross-Domain Crowd Counting via Knowledge Diffusion
    Xie, Haiyang
    Yang, Zhengwei
    Zhu, Huilin
    Wang, Zheng
    PROCEEDINGS OF THE 31ST ACM INTERNATIONAL CONFERENCE ON MULTIMEDIA, MM 2023, 2023, : 6520 - 6529