Deep neural networks-based relevant latent representation learning for hyperspectral image classification

作者：

Highlights：

• Learning relevant feature in HSI using multi-view deep autoencoder model.

• Propose a semi-supervised GCN for HSI classification.

• Preserve deep spectral-spatial relevant features in the classification.

• Improve image classification compared to well established DL techniques.

摘要

•Learning relevant feature in HSI using multi-view deep autoencoder model.•Propose a semi-supervised GCN for HSI classification.•Preserve deep spectral-spatial relevant features in the classification.•Improve image classification compared to well established DL techniques.

论文关键词：Deep learning,Representation learning,Hyperspectral image classification,Feature extraction

论文评审过程：Received 15 March 2021, Revised 24 June 2021, Accepted 2 August 2021, Available online 3 August 2021, Version of Record 16 August 2021.

论文官网地址：https://doi.org/10.1016/j.patcog.2021.108224