TY - GEN
T1 - HYPERSPECTRAL CLASSIFICATION USING COOPERATIVE SPATIAL-SPECTRAL ATTENTION NETWORK WITH TENSOR LOW-RANK RECONSTRUCTION
AU - Li, Sen
AU - Luo, Xiaoyan
AU - Wang, Qixiong
AU - Li, Lei
AU - Shen, Weifa
AU - Yin, Jihao
N1 - Publisher Copyright:
© 2021 IEEE.
PY - 2021
Y1 - 2021
N2 - Spatial and spectral attention networks have been both well introduced to Hyperspectral image (HSI) classification. However, in previous works, they are seldom considered jointly. To obtain a 3D spatial-spectral attention map, which is beneficial for extracting discriminative spatial-spectral features, we propose a novel cooperative spatial-spectral attention network with tensor low-rank reconstruction. Firstly, a tensor low-rank reconstruction (TLRR) block is designed to learn a spatial-spectral attention map tensor, which adaptively emphasizes the attention features of the salient spatial positions and informative spectral bands simultaneously. Secondly, these attention features are merged into simple convolutional features which are more discriminative for classification. Finally, the experimental results demonstrate that our proposed method outperforms some state-of-the-art methods on two typical HSI datasets.
AB - Spatial and spectral attention networks have been both well introduced to Hyperspectral image (HSI) classification. However, in previous works, they are seldom considered jointly. To obtain a 3D spatial-spectral attention map, which is beneficial for extracting discriminative spatial-spectral features, we propose a novel cooperative spatial-spectral attention network with tensor low-rank reconstruction. Firstly, a tensor low-rank reconstruction (TLRR) block is designed to learn a spatial-spectral attention map tensor, which adaptively emphasizes the attention features of the salient spatial positions and informative spectral bands simultaneously. Secondly, these attention features are merged into simple convolutional features which are more discriminative for classification. Finally, the experimental results demonstrate that our proposed method outperforms some state-of-the-art methods on two typical HSI datasets.
KW - Attention mechanism
KW - Convolutional neural network (CNN)
KW - Hyperspectral image (HSI) classification
KW - Low-rank reconstruction
UR - https://www.scopus.com/pages/publications/85125564692
U2 - 10.1109/ICIP42928.2021.9506400
DO - 10.1109/ICIP42928.2021.9506400
M3 - 会议稿件
AN - SCOPUS:85125564692
T3 - Proceedings - International Conference on Image Processing, ICIP
SP - 2658
EP - 2662
BT - 2021 IEEE International Conference on Image Processing, ICIP 2021 - Proceedings
PB - IEEE Computer Society
T2 - 28th IEEE International Conference on Image Processing, ICIP 2021
Y2 - 19 September 2021 through 22 September 2021
ER -