Few-Shot SAR Target Recognition via Causal Inference and Deep Metric Learning

Deep learning, with large-scale annotated datasets, has demonstrated remarkable success in synthetic aperture radar automatic target recognition (SAR-ATR). However, the collecting of SAR images is expensive and complex, and manually labeling them requires expert knowledge. To overcome these limitati...

Full description

Saved in:
Bibliographic Details
Main Authors: Ke Wang, Yuqian Mao, Qi Qiao
Format: Article
Language:English
Published: IEEE 2025-01-01
Series:IEEE Access
Subjects:
Online Access:https://ieeexplore.ieee.org/document/11080412/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850077895078182912
author Ke Wang
Yuqian Mao
Qi Qiao
author_facet Ke Wang
Yuqian Mao
Qi Qiao
author_sort Ke Wang
collection DOAJ
description Deep learning, with large-scale annotated datasets, has demonstrated remarkable success in synthetic aperture radar automatic target recognition (SAR-ATR). However, the collecting of SAR images is expensive and complex, and manually labeling them requires expert knowledge. To overcome these limitations, we propose a few-shot learning model capable of accurate recognition of novel targets with minimal training samples. Our model innovatively integrates causal inference with mutual centralized learning (MCL) to address few-shot SAR-ATR tasks. First, we establish a causal inference framework to identify and model the dependencies among target characteristics, imaging conditions, and category labels. This framework incorporates a novel causal intervention method based on multi-scale random convolution to eliminate spurious correlations caused by imaging variations, thereby enhancing feature stability. Second, we introduce an advanced MCL module to effectively evaluate feature similarity in few-shot settings. MCL breaks through the unidirectional matching paradigm adopted by conventional metric learning. Through its bidirectional feature interactions and dense feature accessibility mechanisms, MCL achieves more robust feature discrimination in few-shot learning tasks. Comprehensive experiments demonstrate that our model outperforms existing advanced few-shot SAR-ATR methods, achieving superior recognition accuracy while maintaining robustness in data-scarce scenarios.
format Article
id doaj-art-d9bac4a12ea640e1a9f8e28cade6e184
institution DOAJ
issn 2169-3536
language English
publishDate 2025-01-01
publisher IEEE
record_format Article
series IEEE Access
spelling doaj-art-d9bac4a12ea640e1a9f8e28cade6e1842025-08-20T02:45:42ZengIEEEIEEE Access2169-35362025-01-011312498812500210.1109/ACCESS.2025.358919211080412Few-Shot SAR Target Recognition via Causal Inference and Deep Metric LearningKe Wang0https://orcid.org/0000-0003-2982-4153Yuqian Mao1https://orcid.org/0000-0002-1615-4165Qi Qiao2https://orcid.org/0009-0000-2334-8956School of Computer and Communication, Jiangsu Vocational College of Electronics and Information, Huaian, ChinaThe Eighth Research Institute of China, China State Shipbuilding Corporation Ltd., Nanjing, ChinaSchool of Computer and Communication, Jiangsu Vocational College of Electronics and Information, Huaian, ChinaDeep learning, with large-scale annotated datasets, has demonstrated remarkable success in synthetic aperture radar automatic target recognition (SAR-ATR). However, the collecting of SAR images is expensive and complex, and manually labeling them requires expert knowledge. To overcome these limitations, we propose a few-shot learning model capable of accurate recognition of novel targets with minimal training samples. Our model innovatively integrates causal inference with mutual centralized learning (MCL) to address few-shot SAR-ATR tasks. First, we establish a causal inference framework to identify and model the dependencies among target characteristics, imaging conditions, and category labels. This framework incorporates a novel causal intervention method based on multi-scale random convolution to eliminate spurious correlations caused by imaging variations, thereby enhancing feature stability. Second, we introduce an advanced MCL module to effectively evaluate feature similarity in few-shot settings. MCL breaks through the unidirectional matching paradigm adopted by conventional metric learning. Through its bidirectional feature interactions and dense feature accessibility mechanisms, MCL achieves more robust feature discrimination in few-shot learning tasks. Comprehensive experiments demonstrate that our model outperforms existing advanced few-shot SAR-ATR methods, achieving superior recognition accuracy while maintaining robustness in data-scarce scenarios.https://ieeexplore.ieee.org/document/11080412/Synthetic aperture radar (SAR)target recognitiondeep metric learningcausal inference
spellingShingle Ke Wang
Yuqian Mao
Qi Qiao
Few-Shot SAR Target Recognition via Causal Inference and Deep Metric Learning
IEEE Access
Synthetic aperture radar (SAR)
target recognition
deep metric learning
causal inference
title Few-Shot SAR Target Recognition via Causal Inference and Deep Metric Learning
title_full Few-Shot SAR Target Recognition via Causal Inference and Deep Metric Learning
title_fullStr Few-Shot SAR Target Recognition via Causal Inference and Deep Metric Learning
title_full_unstemmed Few-Shot SAR Target Recognition via Causal Inference and Deep Metric Learning
title_short Few-Shot SAR Target Recognition via Causal Inference and Deep Metric Learning
title_sort few shot sar target recognition via causal inference and deep metric learning
topic Synthetic aperture radar (SAR)
target recognition
deep metric learning
causal inference
url https://ieeexplore.ieee.org/document/11080412/
work_keys_str_mv AT kewang fewshotsartargetrecognitionviacausalinferenceanddeepmetriclearning
AT yuqianmao fewshotsartargetrecognitionviacausalinferenceanddeepmetriclearning
AT qiqiao fewshotsartargetrecognitionviacausalinferenceanddeepmetriclearning