Target Classification Using the Deep Convolutional Networks for SAR Images

By Sizhe Chen, Haipeng Wang, Feng Xu, Yaqiu Jin. IEEE Transactions on Geoscience and Remote Sensing, 2016

1,187 citations recorded by the source corpus.

Subfields: Aerospace Engineering, Ocean Engineering, Oceanography

View Paper: https://www.semanticscholar.org/paper/2d92eac985bd87556a414f7bd007d98af8a03924

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Aerospace Engineering

1. Mohammad Rostami
2. Xiaoxiang Zhu
3. Jianwei Ma
4. Siwei Yu
5. Richard Bamler
6. Lichao Mou
7. Feng Xu
8. Zhimian Zhang

In Ocean Engineering

1. Mohammad Rostami
2. Siwei Yu
3. Jianwei Ma
4. Lichao Mou
5. Gui-Song Xia
6. Feng Xu
7. Haipeng Wang
8. Yaqiu Jin

In Oceanography

1. Mohammad Rostami
2. Jianwei Ma
3. Siwei Yu
4. Feng Xu
5. Haipeng Wang
6. Yaqiu Jin
7. Serban Stan

Citations in the corpus

• Deep learning in remote sensing: a review (2017, background): “[42] propose to further remove all fully connected layers from conventional CNNs, which are accountable for most trainable parameters.”
• Fusion of Deep Convolutional Neural Networks for Land Cover Classification of High-Resolution Imagery (2017, methodology): “DCNN has been used to classify a wide variety of remotely sensed imagery, including synthetic aperture radar [1], hyperspectral [2], and high-resolution electro-optical imagery [3]–[5].”
• SAR Image Classification Using Few-Shot Cross-Domain Transfer Learning (2019, methodology): “Additionally, training a CNN using most existing SAR datasets leads to overfit models that do not generalize well on test sets [3, 32].”
• Class Boundary Exemplar Selection Based Incremental Learning for Automatic Target Recognition (2020, methodology): “Most recent methods for ATR are based on supervised learning [1], [2].”
• Deep Learning for Geophysics: Current and Future Trends (2021, methodology): “The first introduction of DL in subjects related to geophysics focused on remote sensing in 2016 and 2017 (Chen, Jiang, et al., 2016; Chen, Wang, et al., 2016; Maggiori et al., 2017; Li et al., 2017), since remote sensing is a common technique widely used in many areas.”
• SAR Automatic Target Recognition Based on Multiview Deep Learning Framework (2018, background): “Because of the superiority of deep neural networks, some significant works have been done to design deep learning architectures in SAR ATR with outstanding recognition results [14], [23], [24].”
• Fine-Grained Continual Learning for SAR Target Recognition (2022, background): “For instance, in [1], a new all-convolutional networks with sparse-”
• Polarimetric SAR Image Classification Using Deep Convolutional Neural Networks (2016, methodology): “The model was trained using back-propagation and stochastic gradient descent with minibatch size of 64 [14].”
• OpenSARShip: A Dataset Dedicated to Sentinel-1 Ship Interpretation (2018, background): “Moving and stationary target dataset [16] composed of ten types of ground targets with different aspect angles, depression angles, and target configurations is widely used [17]–[19]; however, this dataset is applicable for ground object extraction and automatic target recognition [18], [19] rather t”
• Deep Transfer Learning for Few-Shot SAR Image Classification (2019, background): “79 In particular, given the size of most existing SAR datasets, training a CNN leads to overfit models as the number 80 of data points are considerably less than the required sample complexity of training a deep network [8,9].”
• Automatic Target Recognition in Synthetic Aperture Radar Imagery: A State-of-the-Art Review (2016, background): “Finally, another recent development is the utilization of the so-called all-convolutional networks (A-ConvNets) [107].”
• Transfer Learning with Deep Convolutional Neural Network for SAR Target Classification with Limited Labeled Data (2017, methodology): “Recently, in [24], a five-layer all-convolutional network was proposed.”
• Complex-Valued Convolutional Neural Network and Its Application in Polarimetric SAR Image Classification (2017, methodology): “[7] adopted a deep convolutional network for SAR target identification and achieved a record-breaking accuracy of over 99% for MSTAR 10-class data sets.”
• Deep Learning Meets SAR: Concepts, models, pitfalls, and perspectives (2020, methodology): “[88] proposed A-ConvNets, a simple 5-layer CNN that was able to achieve state-of-the-art accuracy of about 99% on the MSTAR dataset.”
• Squeeze and Excitation Rank Faster R-CNN for Ship Detection in SAR Images (2019, background): “Due to a successful attempt in classification [6], the deep learning becomes a new choice.”
• Learning Transferable Knowledge Through Embedding Spaces (2019, background): “In particular, given the size of most existing SAR datasets, training a CNN leads to overfitted models as the number of data points are considerably less than the required sample complexity of training a deep network [152; 153].”