Table 1 Studies on signature-band extraction or selection with deep learning methods.

From: Hyperspectral signature-band extraction and learning: an example of sugar content prediction of Syzygium samarangense

Reference

Method

Target

Zhan et al.19

Occlusion based method

Classification of Indian Pines data sets

Darling et al.20

Produced contribution matrix by trained weight matrix

Classification of U.S. Army ERDC data sets

Mou et al.21

Deep reinforcement learning

Classification of Pavia University, Botswana, Indian Pines, and MUUFL Gulfport data sets

Elkholy et al.22

Unsupervised learning with deep autoencoder unmixing

Classification of Kennedy Space Center data sets

Cai et al.23

Contribution map-based CNN

Classification of Pavia University and Indian Pines data sets

This study

Absolute mean of the integrated gradients method

Regression on multispectral data sets