ENGLISH

SAR ship detection in complex background based on multi-feature fusion and non-local channel attention mechanism

Book information

Publisher
Taylor and Francis
Year
2021
Language
english
Format
PDF
Filesize
22 MB (23467919 bytes)
Pages
7519-7550\33
Time added
2023-02-06 14:18:03

Description

Abstract 1. Introduction 2. Proposed method 2.1. Single shot multibox detector algorithm 2.2. MFNL-Net model 2.3. Non-local channel attention block 2.4. Multi-scale spatial pyramid pooling 2.5. Bounding boxes matching and loss function 3. Experiments and analysis 3.1. Experiment Dataset 3.2. Experimental settings 3.3. Evaluation Metrics 3.4. SAR ship feature extraction 3.5. The effect of NLCA-Block 3.6. The effect of MSPP 3.7. Experiments on SSDD Dataset 3.8. Experiments on GS-Ship Dataset 3.9. Experiments for Different Resolution SAR Images 3.10. Robustness analysis 3.11. Analysis on Failure Cases and False Alarms 4. Conclusion Acknowledgements Disclosure statement ORCID References

Similar books