Deep Induction Network for Small Samples Classification of Hyperspectral Images

被引:34
作者
Gao, Kuiliang [1 ]
Guo, Wenyue [1 ]
Yu, Xuchu [1 ]
Liu, Bing [1 ]
Yu, Anzhu [1 ]
Wei, Xiangpo [1 ]
机构
[1] PLA Strateg Support Force Informat Engn Univ, Zhengzhou 450001, Peoples R China
基金
中国国家自然科学基金;
关键词
Feature extraction; Deep learning; Training; Convolution; Task analysis; Hyperspectral imaging; hyperspectral images (HSI) classification; induction network; meta-learning; small samples classification (SSC); SPECTRAL-SPATIAL CLASSIFICATION; FEATURE-EXTRACTION; CLASSIFIERS;
D O I
10.1109/JSTARS.2020.3002787
中图分类号
TM [电工技术]; TN [电子技术、通信技术];
学科分类号
0808 ; 0809 ;
摘要
Recently, the deep learning models have achieved great success in hyperspectral images (HSI) classification. However, most of the deep learning models fail to obtain satisfactory results under the condition of small samples due to the contradiction between the large parameter space of the deep learning models and the insufficient labeled samples in HSI. To address the problem, a deep model based on the induction network is designed in this article to improve the classification performance of HSI under the condition of small samples. Specifically, the typical meta-training strategy is adopted, enabling the model to acquire stronger generalization ability, so as to accurately distinguish the new classes with only a few labeled samples (e.g., five samples per class). Moreover, in order to deal with the disturbance caused by the various characteristics of the samples in the same class in HSI, the class-wise induction module is introduced utilizing the dynamic routing algorithm, which can induce the sample-wise representations to the class-wise level representations. The obtained class-wise level representations possess better separability, allowing the designed model to generate more accurate and robust classification results. Extensive experiments are carried out on three public HSI to verify the effectiveness of the proposed method. The results demonstrate that our method outperforms existing deep learning methods under the condition of small samples.
引用
收藏
页码:3462 / 3477
页数:16
相关论文
共 57 条
[1]  
Agarwal A, 2007, 2007 IEEE INTERNATIONAL SYMPOSIUM ON SIGNAL PROCESSING AND INFORMATION TECHNOLOGY, VOLS 1-3, P1077
[2]  
Bertinetto L., 2019, INT C LEARNING REPRE
[3]   Hyperspectral imagery classification with cascaded support vector machines and multi-scale superpixel segmentation [J].
Cao, Xianghai ;
Wang, Da ;
Wang, Xiaozhen ;
Zhao, Jing ;
Jiao, Licheng .
INTERNATIONAL JOURNAL OF REMOTE SENSING, 2020, 41 (12) :4528-4548
[4]   Deep Feature Extraction and Classification of Hyperspectral Images Based on Convolutional Neural Networks [J].
Chen, Yushi ;
Jiang, Hanlu ;
Li, Chunyang ;
Jia, Xiuping ;
Ghamisi, Pedram .
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, 2016, 54 (10) :6232-6251
[5]   Deep Learning-Based Classification of Hyperspectral Data [J].
Chen, Yushi ;
Lin, Zhouhan ;
Zhao, Xing ;
Wang, Gang ;
Gu, Yanfeng .
IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING, 2014, 7 (06) :2094-2107
[6]   Hyperspectral Images Classification Based on Dense Convolutional Networks with Spectral-Wise Attention Mechanism [J].
Fang, Bei ;
Li, Ying ;
Zhang, Haokui ;
Chan, Jonathan Cheung-Wai .
REMOTE SENSING, 2019, 11 (02)
[7]  
Finn C, 2017, PR MACH LEARN RES, V70
[8]   Deep Relation Network for Hyperspectral Image Few-Shot Classification [J].
Gao, Kuiliang ;
Liu, Bing ;
Yu, Xuchu ;
Qin, Jinchun ;
Zhang, Pengqiang ;
Tan, Xiong .
REMOTE SENSING, 2020, 12 (06)
[9]  
Geng Ruiying, 2019, ARXIV190210482
[10]   Advanced Spectral Classifiers for Hyperspectral Images A review [J].
Ghamisi, Pedram ;
Plaza, Javier ;
Chen, Yushi ;
Li, Jun ;
Plaza, Antonio .
IEEE GEOSCIENCE AND REMOTE SENSING MAGAZINE, 2017, 5 (01) :8-32