Abstract

Band selection is a commonly used approach for dimensionality reduction in hyperspectral imagery. Affinity propagation (AP), a new clustering algorithm, is addressed in many fields, and it can be used for hyperspectral band selection. However, this algorithm cannot get a fixed number of exemplars during the message-passing procedure, which limits its uses to a great extent. This paper proposes an adaptive AP (AAP) algorithm for semi-supervised hyperspectral band selection and investigates the effectiveness of distance metrics for improving band selection. Specifically, the exemplar number determination algorithm and bisection method are addressed to improve AP procedure, and the relations between selected exemplar numbers and preferences are established. Experiments are conducted to evaluate the proposed AAP-based band selection algorithm, and the results demonstrate that the proposed method outperforms other popular methods, with lower computational cost and robust results.

©2012 Optical Society of America

Self-training-based face recognition using semi-supervised linear discriminant analysis and affinity propagation

Haitao Gan, Nong Sang, and Rui Huang
J. Opt. Soc. Am. A 31(1) 1-6 (2014)

Hyperspectral band selection based on a variable precision neighborhood rough set

Yao Liu, Hong Xie, Liguo Wang, and Kezhu Tan
Appl. Opt. 55(3) 462-472 (2016)

Selection of optimal combinations of band-pass filters for ice detection by hyperspectral imaging

Shigeki Nakauchi, Ken Nishino, and Takuya Yamashita
Opt. Express 20(2) 986-1000 (2012)

Advanced similarity metric and its application to the band selection issue of multiple spectra identification

Chunhui Zhao, Minghua Tian, Wu Liu, Xifeng Yao, and Bin Qi
Appl. Opt. 55(21) 5790-5805 (2016)

Stochastic spectral unmixing with enhanced endmember class separation

Michael T. Eismann and Russell C. Hardie
Appl. Opt. 43(36) 6596-6608 (2004)