Journal
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
Volume 59, Issue 3, Pages 2430-2447Publisher
IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
DOI: 10.1109/TGRS.2020.3005431
Keywords
Attention mechanism; deep learning (DL); hyperspectral image (HSI) classification; information complement; spatial-spectral features
Categories
Funding
- National Natural Science Foundation of China [61801351, 61802190, 61772400, 61672444, 61272366]
- Key Laboratory of National Defense Science and Technology Foundation Project [6142113180302]
- China Post-Doctoral Science Foundation [2017M620441]
- Xidian University New Teacher Innovation Fund Project [XJS18032]
- Hong Kong Baptist University (HKBU), Research Committee, Initiation Grant-Faculty Niche Research Areas (IG-FNRA) 2018/19 [RC-FNRA-IG/1819/SCI/03]
- Innovation and Technology Fund of Innovation and Technology Commission of the Government of the Hong Kong SAR [ITS/339/18]
- Faculty Research Grant of HKBU [FRG2/17-18/082]
- Shenzhen Science, Technology and Innovation Commission (SZSTI) [JCYJ20160531194006833]
Ask authors/readers for more resources
In this article, a 3-D octave convolution with spatial-spectral attention network (3DOC-SSAN) is proposed to improve the classification accuracy of hyperspectral images (HSIs). By integrating spatial and spectral information through an information complement model, important spatial and spectral features can be fully utilized for classification tasks. Comparing with existing classifiers, this method shows competitive performance on benchmark datasets.
In recent years, with the development of deep learning (DL), the hyperspectral image (HSI) classification methods based on DL have shown superior performance. Although these DL-based methods have great successes, there is still room to improve their ability to explore spatial-spectral information. In this article, we propose a 3-D octave convolution with the spatial-spectral attention network (3DOC-SSAN) to capture discriminative spatial-spectral features for the classification of HSIs. Especially, we first extend the octave convolution model using 3-D convolution, namely, a 3-D octave convolution model (3D-OCM), in which four 3-D octave convolution blocks are combined to capture spatial-spectral features from HSIs. Not only the spatial information can be mined deeply from the highand low-frequency aspects but also the spectral information can be taken into account by our 3D-OCM. Second, we introduce two attention models from spatial and spectral dimensions to highlight the important spatial areas and specific spectral bands that consist of significant information for the classification tasks. Finally, in order to integrate spatial and spectral information, we design an information complement model to transmit important information between spatial and spectral attention features. Through the information complement model, the beneficial parts of spatial and spectral attention features for the classification tasks can be fully utilized. Comparing with several existing popular classifiers, our proposed method can achieve competitive performance on four benchmark data sets.
Authors
I am an author on this paper
Click your name to claim this paper and add it to your profile.
Reviews
Recommended
No Data Available