University of Leicester
Browse
final version.pdf (2.93 MB)

SAR Target Incremental Recognition based on Features with Strong Separability

Download (2.93 MB)
journal contribution
posted on 2024-01-17, 14:14 authored by F Gao, L Kong, R Lang, J Sun, J Wang, A Hussain, Huiyu Zhou

With the rapid development of deep learning technology, many SAR target recognition algorithms based on convolutional neural networks have achieved exceptional performance on various datasets. However, conventional neural networks are repeatedly iterated on a fixed dataset until convergence, and once they learn new tasks, a large amount of previously learned knowledge is forgotten, leading to a significant decline in performance on old tasks. This paper presents an incremental learning method based on strong separability features (SSF-IL) to address the model’s forgetting of previously learned knowledge. The SSF-IL employs both intra-class and inter-class scatter to compute the feature separability loss, in order to enhance the linear separability of features during incremental learning. In the process of learning new classes, an intra-class clustering loss is proposed to replace the conventional knowledge distillation. This loss function constrains the old class features to cluster around the saved class centers, maintaining the separability among the old class features. Finally, a classifier bias correction method based on boundary features is designed to reinforce the classifier’s decision boundary and reduce classification errors. SAR target incremental recognition experiments are conducted on the MSTAR dataset, and the results are compared with several existing incremental learning algorithms to demonstrate the effectiveness of the proposed algorithm.

Funding

Towards visually-driven speech enhancement for cognitively-inspired multi-modal hearing-aid devices (AV-COGHEAR)

Engineering and Physical Sciences Research Council

Find out more...

COG-MHEAR: Towards cognitively-inspired 5G-IoT enabled, multi-modal Hearing Aids

Engineering and Physical Sciences Research Council

Find out more...

Natural Language Generation for Low-resource Domains

Engineering and Physical Sciences Research Council

Find out more...

National Natural Science Foundation of China under Grant 62371022

History

Author affiliation

Department of Informatics, University of Leicester

Version

  • AM (Accepted Manuscript)

Published in

IEEE Transactions on Geoscience and Remote Sensing

Publisher

Institute of Electrical and Electronics Engineers

issn

0196-2892

Copyright date

2024

Available date

2024-01-17

Language

en

Usage metrics

    University of Leicester Publications

    Categories

    No categories selected

    Licence

    Exports

    RefWorks
    BibTeX
    Ref. manager
    Endnote
    DataCite
    NLM
    DC