An attention-driven convolutional neural network-based multi-level spectral–spatial feature learning for hyperspectral image classification

作者：

Highlights：

• The proposed MFCNN aggregates multiple adjacent backbones to extract features.

• A SSAM learns the multi-scale correlations between spectral–spatial features.

• The SSAM method can recalibrate the discriminative features learned by MFCNN.

• A MFNSAM combines SSAM with MFCNN for hyperspectral image classification.

摘要

•The proposed MFCNN aggregates multiple adjacent backbones to extract features.•A SSAM learns the multi-scale correlations between spectral–spatial features.•The SSAM method can recalibrate the discriminative features learned by MFCNN.•A MFNSAM combines SSAM with MFCNN for hyperspectral image classification.

论文关键词：Hyperspectral image classification,Feature extraction,Convolutional neural networks,Attention mechanism,Spectral–spatial feature

论文评审过程：Received 12 March 2021, Revised 22 July 2021, Accepted 23 July 2021, Available online 31 July 2021, Version of Record 4 August 2021.

论文官网地址：https://doi.org/10.1016/j.eswa.2021.115663