A Distribution and Structure Match Generative Adversarial Network for SAR Image Classification
A Distribution and Structure Match Generative Adversarial Network for SAR Image Classification
复制标题
用于 SAR 图像分类的分布和结构匹配生成对抗网络
DOI：
10.1109/tgrs.2019.2959120
复制
发表时间：
2020-01
影响因子：
8.2
通讯作者：
Jiao Licheng
中科院分区：
文献类型：
--
作者：
Ren Zhongle;Hou Biao;Wu Qian;Wen Zaidao;Jiao Licheng
合成孔径雷达（SAR）图像分类是SAR图像解译的基础研究。以往的方法都是基于统计特征或空间特征的，无法提取具有完整SAR图像特征的特征，从而限制了分类的性能。在这篇文章中，提出了新的样本加权和类对抗训练策略，融合互补的SAR特征。在此基础上，构造了一个分布和结构匹配的辅助分类器生成对抗网络（DSM-ACGAN），用于高质量的鉴别特征学习。特别地，在DSM-ACGAN的类对抗训练中，统计分布和空间结构的特征被联合考虑。一方面，DSM-ACGAN将SAR图像的真实特征作为目标，让生成器学习每一类的生成模型。另一方面，更重要的是，它指导了机器人同时捕获所需的统计和结构特征。通过类对抗处理，区分性特征学习逐步改进并有助于分类。此外，可以生成类平衡和合理的样本。对三幅不同卫星宽SAR图像的实验结果证实了DSM-ACGAN中类对抗训练的有效性和判别特征学习的优越性。视觉性能和定量指标也显示了新模型的最新性能。
Synthetic aperture radar (SAR) image classification is a fundamental research in the interpretation of SAR images. The previous methods are unilaterally based on statistical features or spatial features, which cannot capture features with complete SAR image characteristics and unavoidably limits the performance for classification. In this article, novel sample weighting and class adversarial training strategies are proposed to fuse complementary SAR characteristics. Based on these, a distribution and structure match auxiliary classifier generative adversarial network (DSM-ACGAN) is constructed for high-quality discriminative feature learning. Particularly, the characteristics of statistical distribution and spatial structure are jointly considered in class adversarial training of DSM-ACGAN. On the one hand, DSM-ACGAN sets the true SAR image characteristics as goals for the generator to learn generative models of each category. On the other hand, and more importantly, it guides the discriminator to simultaneously capture the desired statistical and structural features. Through the class adversarial processing, the discriminative feature learning progressively improves and contributes to classification. Additionally, class-balanced and plausible samples can be generated. Experimental results on three broad SAR images from different satellites confirm the effectiveness of class adversarial training and the superiority of discriminative feature learning in DSM-ACGAN. Visual performance and quantitative metrics also show the state-of-the-art performance of the novel model.
登录
查看更多内容
用于高光谱图像分类的可变形卷积神经网络
影响因子：
4.8
作者：
Jian Zhu;Leyuan Fang;Pedram Ghamisi
通讯作者：
Pedram Ghamisi
DOI：
10.1109/tgrs.2018.2794326
发表时间：
2018-06-01
影响因子：
8.2
作者：
Song, Weiwei;Li, Shutao;Lu, Ting
通讯作者：
Lu, Ting
DOI：
10.1109/jstars.2018.2803212
发表时间：
2018-03
影响因子：
5.5
作者：
N. Merkle;S. Auer;R. Müller;P. Reinartz
通讯作者：
N. Merkle;S. Auer;R. Müller;P. Reinartz
影响因子：
4.8
作者：
V. Krylov;G. Moser;S. Serpico;J. Zerubia
通讯作者：
V. Krylov;G. Moser;S. Serpico;J. Zerubia
影响因子：
4.8
作者：
E. Uslu;S. Albayrak
通讯作者：
E. Uslu;S. Albayrak