SCF-CIL: A Multi-Stage Regularization-Based SAR Class-Incremental Learning Method Fused with Electromagnetic Scattering Features

Synthetic aperture radar (SAR) recognition systems often need to collect new data and update the network accordingly. However, the network faces the challenge of catastrophic forgetting, where previously learned knowledge might be lost during the incremental learning of new data. To improve the appl...

Full description

Saved in:
Bibliographic Details
Main Authors: Yunpeng Zhang, Mengdao Xing, Jinsong Zhang, Sergio Vitale
Format: Article
Language:English
Published: MDPI AG 2025-04-01
Series:Remote Sensing
Subjects:
Online Access:https://www.mdpi.com/2072-4292/17/9/1586
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850277594281279488
author Yunpeng Zhang
Mengdao Xing
Jinsong Zhang
Sergio Vitale
author_facet Yunpeng Zhang
Mengdao Xing
Jinsong Zhang
Sergio Vitale
author_sort Yunpeng Zhang
collection DOAJ
description Synthetic aperture radar (SAR) recognition systems often need to collect new data and update the network accordingly. However, the network faces the challenge of catastrophic forgetting, where previously learned knowledge might be lost during the incremental learning of new data. To improve the applicability and sustainability of SAR target classification methods, we propose a multi-stage regularization-based class-incremental learning (CIL) method for SAR targets, called SCF-CIL, which addresses catastrophic forgetting. This method offers three main contributions. First, for the feature extractor, we fuse the convolutional neural network features with the scattering center features using a cross-attention feature fusion structure, ensuring both the plasticity and stability of the extracted features. Next, an overfitting training strategy is applied to provide clustering space for unseen classes with an acceptable trade-off in the accuracy of the current classes. Finally, we analyze the influence of training with imbalanced data on the last fully connected layer and introduce a multi-stage regularization method by dividing the calculation of the fully connected layer into three parts and applying regularization to each. Our experiments on SAR datasets demonstrate the effectiveness of these improvements.
format Article
id doaj-art-04089685f4ce47068f2d24d4152ddcdf
institution OA Journals
issn 2072-4292
language English
publishDate 2025-04-01
publisher MDPI AG
record_format Article
series Remote Sensing
spelling doaj-art-04089685f4ce47068f2d24d4152ddcdf2025-08-20T01:49:50ZengMDPI AGRemote Sensing2072-42922025-04-01179158610.3390/rs17091586SCF-CIL: A Multi-Stage Regularization-Based SAR Class-Incremental Learning Method Fused with Electromagnetic Scattering FeaturesYunpeng Zhang0Mengdao Xing1Jinsong Zhang2Sergio Vitale3The School of Electronic Engineering, Xidian University, Xi’an 710071, ChinaThe National Laboratory of Radar Signal Processing, Xidian University, Xi’an 710071, ChinaThe Academy of Advanced Interdisciplinary Research, Xidian University, Xi’an 710071, ChinaThe Dipartimento di Ingegneria, University of Naples Parthenope, 80143 Naples, ItalySynthetic aperture radar (SAR) recognition systems often need to collect new data and update the network accordingly. However, the network faces the challenge of catastrophic forgetting, where previously learned knowledge might be lost during the incremental learning of new data. To improve the applicability and sustainability of SAR target classification methods, we propose a multi-stage regularization-based class-incremental learning (CIL) method for SAR targets, called SCF-CIL, which addresses catastrophic forgetting. This method offers three main contributions. First, for the feature extractor, we fuse the convolutional neural network features with the scattering center features using a cross-attention feature fusion structure, ensuring both the plasticity and stability of the extracted features. Next, an overfitting training strategy is applied to provide clustering space for unseen classes with an acceptable trade-off in the accuracy of the current classes. Finally, we analyze the influence of training with imbalanced data on the last fully connected layer and introduce a multi-stage regularization method by dividing the calculation of the fully connected layer into three parts and applying regularization to each. Our experiments on SAR datasets demonstrate the effectiveness of these improvements.https://www.mdpi.com/2072-4292/17/9/1586synthetic aperture radarclass-incremental learningclassificationattributed scattering center feature
spellingShingle Yunpeng Zhang
Mengdao Xing
Jinsong Zhang
Sergio Vitale
SCF-CIL: A Multi-Stage Regularization-Based SAR Class-Incremental Learning Method Fused with Electromagnetic Scattering Features
Remote Sensing
synthetic aperture radar
class-incremental learning
classification
attributed scattering center feature
title SCF-CIL: A Multi-Stage Regularization-Based SAR Class-Incremental Learning Method Fused with Electromagnetic Scattering Features
title_full SCF-CIL: A Multi-Stage Regularization-Based SAR Class-Incremental Learning Method Fused with Electromagnetic Scattering Features
title_fullStr SCF-CIL: A Multi-Stage Regularization-Based SAR Class-Incremental Learning Method Fused with Electromagnetic Scattering Features
title_full_unstemmed SCF-CIL: A Multi-Stage Regularization-Based SAR Class-Incremental Learning Method Fused with Electromagnetic Scattering Features
title_short SCF-CIL: A Multi-Stage Regularization-Based SAR Class-Incremental Learning Method Fused with Electromagnetic Scattering Features
title_sort scf cil a multi stage regularization based sar class incremental learning method fused with electromagnetic scattering features
topic synthetic aperture radar
class-incremental learning
classification
attributed scattering center feature
url https://www.mdpi.com/2072-4292/17/9/1586
work_keys_str_mv AT yunpengzhang scfcilamultistageregularizationbasedsarclassincrementallearningmethodfusedwithelectromagneticscatteringfeatures
AT mengdaoxing scfcilamultistageregularizationbasedsarclassincrementallearningmethodfusedwithelectromagneticscatteringfeatures
AT jinsongzhang scfcilamultistageregularizationbasedsarclassincrementallearningmethodfusedwithelectromagneticscatteringfeatures
AT sergiovitale scfcilamultistageregularizationbasedsarclassincrementallearningmethodfusedwithelectromagneticscatteringfeatures