Abstract:Band selection is a research hotspot in hyperspectral image processing. The continuity of the spectral bands causes the adjacent bands to be highly correlated, and correlation among long-range bands is possible with hundreds of spectral bands. Most existing deep learning methods fail to make full use of the inter-band correlation for band selection. In this paper, a novel dual-graph convolutional network based on band attention and a sparse constraint is proposed for band selection. The network consists of two branches. In the attention branch, band-based dual graphs are constructed to encode the contextual correlation of adjacent bands and the structural correlation of long-range bands into non-Euclidean space. Subsequently, the graph convolution-based band attention mechanism is devised to aggregate the band information in the band-based dual graphs and to generate the attention map for all bands. The band attention map is sparsely constrained and embedded as a mask into the trunk branch. In the trunk branch, sample-based dual graphs are constructed to represent the topological information of the samples in the spectral and spatial domains. Furthermore, a dense graph convolutional network is designed to extract and fuse the spatial–spectral and topological features from the shallow to deep layers for classification. A soft-shifting optimization strategy is implemented by defining a new loss from full bands and selected bands to solve the optimization problem caused by the sparse constraint. In this manner, band selection, feature extraction, and classification can be combined into an end-to-end trainable network. The experimental results on representative hyperspectral image datasets demonstrate the superiority of the proposed method over current state-of-the-art band selection methods.

Network topology analysis: A new method for band selection

Band Selection for Hyperspectral Imagery: A New Approach Based on Complex Networks

A New Network Based Feature-Selection Approach for Hyperspcetral Analysis.

An Unsupervised Band Selection Method Based on Triple Constraints and Attention Network for Hyperspectral Imagery

An Advanced Hyperspectral Band Selection Approach Based on Mutual Information.

Band Selection Based Hyperspectral Unmixing

Unsupervised Band Selection for Hyperspectral Imagery Classification Without Manual Band Removal.

Optimal topological design for distributed estimation over sensor networks

A Geometry-Based Band Selection Approach for Hyperspectral Image Analysis

Attend in Bands: Hyperspectral Band Weighting and Selection for Image Classification

Band Selection for Hyperspectral Imagery Using Affinity Propagation

Anomaly Detection-Oriented Band Selection For Hyperspectral Image

Band Selection for Hyperspecral Imagery Based on Particle Swarm Optimization

Research advances and analysis of hyperspectral remote sensing image band selection

Clustering-based Hyperspectral Band Selection Using Sparse Nonnegative Matrix Factorization

Dual-graph convolutional network based on band attention and sparse constraint for hyperspectral band selection

Band selection for hyperspectral images based on SNR estimation and information measure index

High-Efficiency Hyperspectral Unmixing Based on Band Selection

New Band Selection Approach of Hyperspectral Remote Sensing Image

Hyperspectral Band Selection: A Review

Hyperspectral Image Band Selection Based on Genetic Algorithm