Abstract

Synthetic aperture radar automatic target recognition (SAR ATR) uses computer processing capabilities to infer the classes of the targets without human intervention. For SAR ATR, deep learning gradually emerges as a powerful tool and achieves promising performance. However, it faces serious challenges of how to deal with incremental recognition scenarios. The existing deep learning-based SAR ATR methods usually predefine the total number of recognition classes. In realistic applications, the new tasks/classes will be added continuously. If all old data are stored and mixed with newly added data to update the model, the storage pressure and time consumption make the application infeasible. In this article, the high plastic error correction incremental learning (HPecIL) is proposed to address the model degradation and plasticity decline in the incremental scenario. Multiple optimal models trained on old tasks are used to correct accumulative errors and alleviate model degradation. Moreover, the sharp data distribution shift due to newly added data can also result in the model underperforming. A class-balanced training batch is constructed to deal with the issue of unbalanced data distribution. To make a tradeoff between model stability and model plasticity, low-effect nodes in the model are removed to boost the efficiency of model update. The proposed HPecIL outperforms the other state-of-the-art methods in incremental recognition scenarios. The experimental results demonstrate the effectiveness of the proposed method.

Highlights

  • D UE to the superiorities of all-weather, day-and-night, wide-range, and high-resolution imaging, synthetic aperture radar (SAR) has been widely applied in military reconnaissance, geographic information collection, and change detection [1]

  • SAR automatic target recognition (ATR) uses computer processing capabilities to infer the classes of the targets without human intervention [2]

  • Some incremental learning methods transfer knowledge from where yi,j is the soft target generated by the old model for the sample i of class j and yi,j is the output of the current model for the sample i of class j, N is the number of samples, C is the number of classes, T is the temperature scaling parameter

Read more

Summary

INTRODUCTION

D UE to the superiorities of all-weather, day-and-night, wide-range, and high-resolution imaging, synthetic aperture radar (SAR) has been widely applied in military reconnaissance, geographic information collection, and change detection [1]. Few methods pay attention to the class imbalance issue degraded from the dataset shift The pruning initialization part increases the contribution rate of the old data in the training phase to improve the model plasticity. Our contributions are the following: 1) The proposed knowledge inheritance preserves multiple optimal models trained on old data to correct the accumulative errors. The current model accumulates errors due to the continuous reduction of old category data in the training phase, which results in the model performance deteriorates. Due to the emergence of newly added data in the incremental recognition scenario, the exemplars of old classes only account for a small ratio.

RELATED WORK
Definitions of Terms
Incremental Learning
THE PROPOSED METHOD
Overall Process
Main Components
EXPERIMENTAL EVALUATION
Experimental Setup
Incremental Recognition Performance
CONCLUSION
Full Text
Paper version not known

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call

Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.