A Few-Shot Learning Method for SAR Images Based on Weighted Distance and Feature Fusion
Open Access
- 14 September 2022
- journal article
- research article
- Published by MDPI AG in Remote Sensing
- Vol. 14 (18) , 4583
Abstract
Convolutional Neural Network (CNN) has been widely applied in the field of synthetic aperture radar (SAR) image recognition. Nevertheless, CNN-based recognition methods usually encounter the problem of poor feature representation ability due to insufficient labeled SAR images. In addition, the large inner-class variety and high cross-class similarity of SAR images pose a challenge for classification. To alleviate the problems mentioned above, we propose a novel few-shot learning (FSL) method for SAR image recognition, which is composed of the multi-feature fusion network (MFFN) and the weighted distance classifier (WDC). The MFFN is utilized to extract input images’ features, and the WDC outputs the classification results based on these features. The MFFN is constructed by adding a multi-scale feature fusion module (MsFFM) and a hand-crafted feature insertion module (HcFIM) to a standard CNN. The feature extraction and representation capability can be enhanced by inserting the traditional hand-crafted features as auxiliary features. With the aid of information from different scales of features, targets of the same class can be more easily aggregated. The weight generation module in WDC is designed to generate category-specific weights for query images. The WDC distributes these weights along the corresponding Euclidean distance to tackle the high cross-class similarity problem. In addition, weight generation loss is proposed to improve recognition performance by guiding the weight generation module. Experimental results on the Moving and Stationary Target Acquisition and Recognition (MSTAR) dataset and the Vehicle and Aircraft (VA) dataset demonstrate that our proposed method surpasses several typical FSL methods.Keywords
Funding Information
- National Natural Science Foundation of China (61771027, EP/M026981/1, EP/T021063/1, EP/T024917/1, NA160342, 720325)
- UK Engineering and Physical Sciences Research Council (EPSRC) (61771027, EP/M026981/1, EP/T021063/1, EP/T024917/1, NA160342, 720325)
- Royal Society’s Newton Advanced Fellowship (61771027, EP/M026981/1, EP/T021063/1, EP/T024917/1, NA160342, 720325)
- European Union’s Horizon 2020 Research and Innovation Program (61771027, EP/M026981/1, EP/T021063/1, EP/T024917/1, NA160342, 720325)
This publication has 45 references indexed in Scilit:
- Deep Convolutional Highway Unit Network for SAR Target Classification With Limited Labeled Training DataIEEE Geoscience and Remote Sensing Letters, 2017
- Dual-Branch Deep Convolution Neural Network for Polarimetric SAR Image ClassificationApplied Sciences, 2017
- SAR Target Recognition via Supervised Discriminative Dictionary Learning and Sparse Representation of the SAR-HOG FeatureRemote Sensing, 2016
- Deep Residual Learning for Image Recognition2016
- Convolutional Neural Network With Data Augmentation for SAR Target RecognitionIEEE Geoscience and Remote Sensing Letters, 2016
- Spatial Pyramid Pooling in Deep Convolutional Networks for Visual RecognitionIEEE Transactions on Pattern Analysis and Machine Intelligence, 2015
- Fusion of high spatial resolution optical and polarimetric SAR images for urban land cover classification2014
- Face Description with Local Binary Patterns: Application to Face RecognitionIEEE Transactions on Pattern Analysis and Machine Intelligence, 2006
- Image representation using 2D Gabor waveletsIEEE Transactions on Pattern Analysis and Machine Intelligence, 1996
- Visual pattern recognition by moment invariantsIRE Transactions on Information Theory, 1962