SAR ATR based on dividing CNN into CAE and SNN

Li, Xuan; Li, Chunsheng; Wang, Pengbo; Men, Zhirong; Xu, Huaping

doi:10.1109/apsar.2015.7306296

Cited by 29 publications

(20 citation statements)

References 4 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Naturally, it is vital to design a network structure that is capable of making full use of the limited available data (such as the Moving and Stationary Target Acquisition and Recognition (MSTAR) dataset). Recent studies regarding ATR began to adopt convolutional neural networks (CNNs); this approach provides a powerful tool for ATR in SAR, with significant progress in the past years [14][15][16]. However, CNNs remains unable to overcome the aforementioned challenges, partly because of their high dependency on large data for training an excellent model.…”

Section: Introductionmentioning

confidence: 99%

Multi-Stream Convolutional Neural Network for SAR Automatic Target Recognition

Zhao

Zou

et al. 2018

Remote Sensing

View full text Add to dashboard Cite

Despite the fact that automatic target recognition (ATR) in Synthetic aperture radar (SAR) images has been extensively researched due to its practical use in both military and civil applications, it remains an unsolved problem. The major challenges of ATR in SAR stem from severe data scarcity and great variation of SAR images. Recent work started to adopt convolutional neural networks (CNNs), which, however, remain unable to handle the aforementioned challenges due to their high dependency on large quantities of data. In this paper, we propose a novel deep convolutional learning architecture, called Multi-Stream CNN (MS-CNN), for ATR in SAR by leveraging SAR images from multiple views. Specifically, we deploy a multi-input architecture that fuses information from multiple views of the same target in different aspects; therefore, the elaborated multi-view design of MS-CNN enables it to make full use of limited SAR image data to improve recognition performance. We design a Fourier feature fusion framework derived from kernel approximation based on random Fourier features which allows us to unravel the highly nonlinear relationship between images and classes. More importantly, MS-CNN is qualified with the desired characteristic of easy and quick manoeuvrability in real SAR ATR scenarios, because it only needs to acquire real-time GPS information from airborne SAR to calculate aspect differences used for constructing testing samples. The effectiveness and generalization ability of MS-CNN have been demonstrated by extensive experiments under both the Standard Operating Condition (SOC) and Extended Operating Condition (EOC) on the MSTAR dataset. Experimental results have shown that our proposed MS-CNN can achieve high recognition rates and outperform other state-of-the-art ATR methods.

show abstract

Section: Introductionmentioning

confidence: 99%

Multi-Stream Convolutional Neural Network for SAR Automatic Target Recognition

Zhao

Zou

et al. 2018

Remote Sensing

View full text Add to dashboard Cite

show abstract

“…Due to limited training samples, the SAR ATR networks are usually more shallower than the natural image recognition case. As for the type identification, both shallow CAE and CNN are employed to conduct MSTAR and TerraSAR-X data recognition with sparsely connected convolution architectures and fine-tuning strategies, and achieve high recognition accuracy over 98% [17,21,22]. Lately, various handcrafted features combined with deep learning for precise recognition, like multi-aspect scattering feature, texture feature and so on.…”

Section: Introductionmentioning

confidence: 99%

Slim and Efficient Neural Network Design for Resource-Constrained SAR Target Recognition

et al. 2018

View full text Add to dashboard Cite

Deep convolutional neural networks (CNN) have been recently applied to synthetic aperture radar (SAR) for automatic target recognition (ATR) and have achieved state-of-the-art results with significantly improved recognition performance. However, the training period of deep CNN is long, and the size of the network is huge, sometimes reaching hundreds of megabytes. These two factors of deep CNN hinders its practical implementation and deployment in real-time SAR platforms that are typically resource-constrained. To address this challenge, this paper presents three strategies of network compression and acceleration to decrease computing and memory resource dependencies while maintaining a competitive accuracy. First, we introduce a new weight-based network pruning and adaptive architecture squeezing method to reduce the network storage and the time of inference and training process, meanwhile maintain a balance between compression ratio and classification accuracy. Then we employ weight quantization and coding to compress the network storage space. Due to the fact that the amount of calculation is mainly reflected in the convolution layer, a fast approach for pruned convolutional layers is proposed to reduce the number of multiplication by exploiting the sparsity in the activation inputs and weights. Experimental results show that the convolutional neural networks for SAR-ATR can be compressed by 40 × without loss of accuracy, and the number of multiplication can be reduced by 15 × . Combining these strategies, we can easily load the network in resource-constrained platforms, speed up the inference process to get the results in real-time or even retrain a more suitable network with new image data in a specific situation.

show abstract

“…The experiment of this method on MSTAR database obtains a target recognition accuracy of 90.1% for 3 classes and 84.7% for 10 classes. Similarly, Li et al uses the autoencoder machine to initialize the DCNN [11]. The difference is that fully connected layers are used as SNN to work as a final classifier in [11], which greatly reduces training time of DCNN on the premise of ensuring accuracy.…”

Section: I! Introductionmentioning

confidence: 99%

“…Similarly, Li et al uses the autoencoder machine to initialize the DCNN [11]. The difference is that fully connected layers are used as SNN to work as a final classifier in [11], which greatly reduces training time of DCNN on the premise of ensuring accuracy. Although the accuracy of two methods is not very high, there is less dependence on experience of experts in the process of feature learning.…”

Section: I! Introductionmentioning

confidence: 99%

“…Although the classification accuracy is satisfactory, it would be time-consuming if the neurons make an obvious mistake during the training of DCNN. Compared with error cost function, the Cross-Entropy is applied as the quadratic function in [9], [11], [14], while there is a lack of in-depth optimization. To improve the classification ability of CNN, we attempt to add class separability information to cross-entropy cost function as a regularization term in DCNN+SVM model.…”

Section: I! Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Combining Deep Convolutional Neural Network and SVM to SAR Image Target Recognition

Gao

Huang

Wang

et al. 2017

2017 IEEE International Conference on Internet of Things (iThings) and IEEE Green Computing and Communications (GreenCom) and I

View full text Add to dashboard Cite

This version is available at https://strathprints.strath.ac.uk/64414/ Strathprints is designed to allow users to access the research output of the University of Strathclyde. Unless otherwise explicitly stated on the manuscript, Copyright © and Moral Rights for the papers on this site are retained by the individual authors and/or other copyright owners. Please check the manuscript for details of any other licences that may have been applied. You may not engage in further distribution of the material for any profitmaking activities or any commercial gain. You may freely distribute both the url (https://strathprints.strath.ac.uk/) and the content of this paper for research or private study, educational, or not-for-profit purposes without prior permission or charge.Any correspondence concerning this service should be sent to the

show abstract

SAR ATR based on dividing CNN into CAE and SNN

Cited by 29 publications

References 4 publications

Multi-Stream Convolutional Neural Network for SAR Automatic Target Recognition

Multi-Stream Convolutional Neural Network for SAR Automatic Target Recognition

Slim and Efficient Neural Network Design for Resource-Constrained SAR Target Recognition

Combining Deep Convolutional Neural Network and SVM to SAR Image Target Recognition

Contact Info

Product

Resources

About