Few-Shot SAR Target Classification via Metalearning
- 22 February 2021
- journal article
- research article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 60 (01962892), 1-14
- https://doi.org/10.1109/tgrs.2021.3058249
Abstract
The state-of-the-art deep neural networks have made a great breakthrough in remote sensing image classification. However, the heavy dependence on large-scale data sets limits the application of the deep learning to synthetic aperture radar (SAR) automatic target recognition (ATR) field where the target sample set is generally small. In this work, a metalearning framework named MSAR, consisting of a metalearner and a base-learner, is proposed to solve the sample restriction problem, which can learn a good initialization as well as a proper update strategy. After training, MSAR can implement fast adaptation with a few training images on new tasks. To the best of our knowledge, this is the first study to solve a few-shot SAR target classification via metalearning. In particular, the few-task problem is defined by analyzing the effect of available training classes on the performance of metalearning models. In order to reduce the metalearning difficulties caused by the few-task problem, three transfer-learning methods are employed, which can leverage the prior knowledge from the pretraining phase. Besides, we design a hard task mining method for effective metalearning. Based on the Moving and Stationary Target Acquisition and Recognition (MSTAR) data set, a specialized data set named NIST-SAR is devised to train and evaluate the proposed method. The experiments on NIST-SAR have shown that the proposed method yields better performances with the largest absolute improvements of 1.7% and 2.3% for 1-shot and 5-shot, respectively, over the next best, which indicates that the proposed method is promising and metalearning is a feasible solution for few-shot SAR ATR.Keywords
Funding Information
- National Natural Science Foundation of China (61725105, 41801349)
This publication has 33 references indexed in Scilit:
- Deep convolutional neural networks for ATR from SAR imageryPublished by SPIE-Intl Soc Optical Eng ,2015
- ImageNet Large Scale Visual Recognition ChallengeInternational Journal of Computer Vision, 2015
- SAR Automatic Target Recognition Using Discriminative Graphical ModelsIEEE Transactions on Aerospace and Electronic Systems, 2014
- Transfer LearningPublished by IGI Global ,2010
- Support vector machines for SAR automatic target recognitionIEEE Transactions on Aerospace and Electronic Systems, 2001
- SAR ATR performance using a conditionally Gaussian modelIEEE Transactions on Aerospace and Electronic Systems, 2001
- Performance of 10- and 20-target MSE classifiersIEEE Transactions on Aerospace and Electronic Systems, 2000
- SAR ATR: so what's the problem? An MSTAR perspectivePublished by SPIE-Intl Soc Optical Eng ,1999
- Moving and stationary target acquisition and recognition (MSTAR) model-based automatic target recognition: search technology for a robust ATRPublished by SPIE-Intl Soc Optical Eng ,1998
- Model-based SAR ATR systemPublished by SPIE-Intl Soc Optical Eng ,1996