A Novel Tri-Training Technique for the Semi-Supervised Classification of Hyperspectral Images Based on Regularized Local Discriminant Embedding Feature Extraction

Depin Ou,Qian Du,Kun Tan,Yu Chen,Xue Wang,Jishuai Zhu

doi:10.3390/rs11060654

Abstract

This paper introduces a novel semi-supervised tri-training classification algorithm based on regularized local discriminant embedding (RLDE) for hyperspectral imagery. In this algorithm, the RLDE method is used for optimal feature information extraction, to solve the problems of singular values and over-fitting, which are the main problems in the local discriminant embedding (LDE) and local Fisher discriminant analysis (LFDA) methods. An active learning method is then used to select the most useful and informative samples from the candidate set. In the experiments undertaken in this study, the three base classifiers were multinomial logistic regression (MLR), k-nearest neighbor (KNN), and random forest (RF). To confirm the effectiveness of the proposed RLDE method, experiments were conducted on two real hyperspectral datasets (Airborne Visible/Infrared Imaging Spectrometer (AVIRIS) and Reflective Optics System Imaging Spectrometer (ROSIS)), and the proposed RLDE tri-training algorithm was compared with its counterparts of tri-training alone, LDE, and LFDA. The experiments confirmed that the proposed approach can effectively improve the classification accuracy for hyperspectral imagery.

Highlights

Hyperspectral sensors have hundreds of spectrally contiguous bands, which can provide abundant spectral information [1]
Despite the fact that deep learning based methods have been developed for hyperspectral images (HSIs) classification, including convolutional neural networks (CNNs) [6,7,8], 3D convolutional neural networks (3D-CNNs) [9,10], and long short-term memory (LSTM) networks [11,12], these problems still exist
The local discriminant embedding (LDE) and local Fisher discriminant analysis (LFDA) algorithms have the following shortcomings: (1) when the number of training samples is smaller than the spectral dimension, the singular value problem occurs in the process of solving the projection vector and (2) in attempting to preserve the local difference information, the over-fitting problem occurs [46]

Summary

Introduction

Hyperspectral sensors have hundreds of spectrally contiguous bands, which can provide abundant spectral information [1]. Common nonlinear dimension reduction methods include kernel based approaches [30,31] and manifold learning algorithms [32]. In [34], a new nonlinear dimension reduction method combining a kernel function with Fisher discriminant analysis was used in the classification of HSIs. In [35,36], Song et al proposed models to learn a set of robust hash functions to map the high-dimensional data points into binary hash codes by effectively utilizing the local structural information. In [38], locally linear embedding (LLE) was used to embed data points in a low-dimensional space by finding the optimal linear reconstruction in a small neighborhood He et al [39] subsequently proposed the neighborhood preserving embedding algorithm based on LLE, and regarded the error minimization as the objective function. We use an active learning method to select the unlabeled samples and use ensemble learning to improve the classification result

Spatial Mean Filtering and Feature Extraction

Spatial Mean Filtering

Cooperative Training Strategy Combining Local Features

Experimental Results and Analysis

Data Used in the Experiments

Conclusions