Learning From Natural Images in Few-Shot SAR Target Classification

The intricate imaging attributes of synthetic aperture radar (SAR) present a formidable challenge to the prevailing few-shot target classification. In order to address this issue, we study how to leverage natural images to assist with few-shot SAR learning and propose a model with cross-domain gener...

Full description

Saved in:
Bibliographic Details
Main Authors: Songhao Shi, Xiaodan Wang, Yafei Song
Format: Article
Language:English
Published: IEEE 2025-01-01
Series:IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Subjects:
Online Access:https://ieeexplore.ieee.org/document/10946144/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849316063801507840
author Songhao Shi
Xiaodan Wang
Yafei Song
author_facet Songhao Shi
Xiaodan Wang
Yafei Song
author_sort Songhao Shi
collection DOAJ
description The intricate imaging attributes of synthetic aperture radar (SAR) present a formidable challenge to the prevailing few-shot target classification. In order to address this issue, we study how to leverage natural images to assist with few-shot SAR learning and propose a model with cross-domain generalization ability, named CDFS-SAR. The core of our approach is a dual-branch network based on supervised contrastive learning. This network comprises a global branch for holistic image understanding and a local branch for capturing fine-grained image details. By exploiting the potential semantic consistency between these branches, we facilitate the learning of background-separated prior knowledge, enabling effective cross-domain generalization to SAR images. To further address the challenge of category confusion in SAR images, we introduce an embedding space reconstruction module. This module utilizes supervised contrastive learning to enhance intraclass compactness and interclass divergence of features. In addition, we employ a Brownian distance covariance metrics module to overcome the limitations of conventional metric distances, ensuring the model acquires more precise semantic representations. A novel joint loss function is formulated to optimize cross-domain generalization and accommodate the unique imaging characteristics of SAR. Experimental results on the MSTAR and OpenSARShip datasets demonstrate the effectiveness of our method and highlight the ability to improve few-shot SAR target classification.
format Article
id doaj-art-95bc6b7f7ed148e38738456133d92bb2
institution Kabale University
issn 1939-1404
2151-1535
language English
publishDate 2025-01-01
publisher IEEE
record_format Article
series IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
spelling doaj-art-95bc6b7f7ed148e38738456133d92bb22025-08-20T03:51:58ZengIEEEIEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing1939-14042151-15352025-01-0118105951060710.1109/JSTARS.2025.355659710946144Learning From Natural Images in Few-Shot SAR Target ClassificationSonghao Shi0https://orcid.org/0009-0008-0232-8460Xiaodan Wang1https://orcid.org/0000-0003-2785-9539Yafei Song2https://orcid.org/0000-0003-0962-0671Air Defense and Antimissile School, Air Force Engineering University, Xi'an, ChinaAir Defense and Antimissile School, Air Force Engineering University, Xi'an, ChinaAir Defense and Antimissile School, Air Force Engineering University, Xi'an, ChinaThe intricate imaging attributes of synthetic aperture radar (SAR) present a formidable challenge to the prevailing few-shot target classification. In order to address this issue, we study how to leverage natural images to assist with few-shot SAR learning and propose a model with cross-domain generalization ability, named CDFS-SAR. The core of our approach is a dual-branch network based on supervised contrastive learning. This network comprises a global branch for holistic image understanding and a local branch for capturing fine-grained image details. By exploiting the potential semantic consistency between these branches, we facilitate the learning of background-separated prior knowledge, enabling effective cross-domain generalization to SAR images. To further address the challenge of category confusion in SAR images, we introduce an embedding space reconstruction module. This module utilizes supervised contrastive learning to enhance intraclass compactness and interclass divergence of features. In addition, we employ a Brownian distance covariance metrics module to overcome the limitations of conventional metric distances, ensuring the model acquires more precise semantic representations. A novel joint loss function is formulated to optimize cross-domain generalization and accommodate the unique imaging characteristics of SAR. Experimental results on the MSTAR and OpenSARShip datasets demonstrate the effectiveness of our method and highlight the ability to improve few-shot SAR target classification.https://ieeexplore.ieee.org/document/10946144/Contrastive learningcross-domain few-shot learning (CDFSL)prototypical networksynthetic aperture radar (SAR) target classification
spellingShingle Songhao Shi
Xiaodan Wang
Yafei Song
Learning From Natural Images in Few-Shot SAR Target Classification
IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Contrastive learning
cross-domain few-shot learning (CDFSL)
prototypical network
synthetic aperture radar (SAR) target classification
title Learning From Natural Images in Few-Shot SAR Target Classification
title_full Learning From Natural Images in Few-Shot SAR Target Classification
title_fullStr Learning From Natural Images in Few-Shot SAR Target Classification
title_full_unstemmed Learning From Natural Images in Few-Shot SAR Target Classification
title_short Learning From Natural Images in Few-Shot SAR Target Classification
title_sort learning from natural images in few shot sar target classification
topic Contrastive learning
cross-domain few-shot learning (CDFSL)
prototypical network
synthetic aperture radar (SAR) target classification
url https://ieeexplore.ieee.org/document/10946144/
work_keys_str_mv AT songhaoshi learningfromnaturalimagesinfewshotsartargetclassification
AT xiaodanwang learningfromnaturalimagesinfewshotsartargetclassification
AT yafeisong learningfromnaturalimagesinfewshotsartargetclassification