Unsupervised and Supervised Feature Extraction Methods for Hyperspectral Images Based on Mixtures of Factor Analyzers

Abstract

This paper proposes three feature extraction (FE) methods based on density estimation for hyperspectral images (HSIs). The methods are a mixture of factor analyzers (MFA), deep MFA (DMFA), and supervised MFA (SMFA). The MFA extends the Gaussian mixture model to allow a low-dimensionality representation of the Gaussians. DMFA is a deep version of MFA and consists of a two-layer MFA, i.e, samples from the posterior distribution at the first layer are input to an MFA model at the second layer. SMFA consists of single-layer MFA and exploits labeled information to extract features of HSI effectively. Based on these three FE methods, the paper also proposes a framework that automatically extracts the most important features for classification from an HSI. The overall accuracy of a classifier is used to automatically choose the optimal number of features and hence performs dimensionality reduction (DR) before HSI classification. The performance of MFA, DMFA, and SMFA FE methods are evaluated and compared to five different types of unsupervised and supervised FE methods by using four real HSIs datasets.

Self-Supervised Assisted Semi-Supervised Residual Network for Hyperspectral Image Classification

“…(3) graph-based methods [21][22][23]; e.g., Ding et al [23] proposed a semi-supervised locality-preserving dense graph neural network (GNN) for HSI classification in which autoregressive moving average filters and context-aware learning are integrated. Moreover, the self-supervised learning methods are also applied to the few-shot HSI classification [24][25][26][27]. In [26], a self-supervised contrastive fruitful disymmetrical expanded network is presented for HSI classification.…”

Minimal learning machine in hyperspectral imaging classification

“…Some of the previous classification studies (eg. 24 ) has cleaned the Indian Pines data, by limiting the classes to the largest ones, since there are classes that has only limited amount of samples (see from table 2). It would have been interesting to try this experiment again with Indian Pines data and limited classes.…”

“…With those methods, we can reduce the number of the training samples (R size) and reduce the computing time, without decreasing the accuracy rate. The FE is an important step before using the HS image classification methods 10,26 .…”

“…The high number of dimensions and the amount of the data makes the processing computationally and memory costly. Other challenges are the curse of the dimensionality 8,10 and the redundancy among the samples 11 . Those challenges might have an impact to the classifiers performance and accuracy.…”

Quantification for Food Inspection enabled by Hyperspectral Imaging System and Machine Learning

“…Even on the 0th day, the negative values of LDA component 1 appeared to originate from the fat tissues whose uorescence was similar to that of NADH. The decision or evaluation boundaries of F. I. produced by QDA are shown in Fig.5bas contour plots 22. Information on F. I., averaged over each hyperspectral image by line-scan-type HIS as reference data, and hyperspectral images taken by snapshot-type HIS were obtained for identical meat specimen as a function of storage time.…”