Semi-supervised classification of hyperspectral imagery based on stacked autoencoders

Fu, Qiongying; Yu, Xi; Wei, Xiangpo; Xue, Zhixiang

doi:10.1117/12.2245011

Cited by 10 publications

(8 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Within this perspective, auto-encoders allow to learn a smart compression with minimal information loss, for example more efficient than a standard PCA. Thus [46] and later [47] proposed dimension reduction through cascade of auto-encoders for denoising followed by classification with a simple perceptron.…”

Section: B Spectral Classificationmentioning

confidence: 99%

Deep Learning for Classification of Hyperspectral Data: A Comparative Review

Audebert

Saux

Lefèvre

2019

IEEE Geosci. Remote Sens. Mag.

499

194

View full text Add to dashboard Cite

In recent years, deep learning techniques revolutionized the way remote sensing data are processed. Classification of hyperspectral data is no exception to the rule, but has intrinsic specificities which make application of deep learning less straightforward than with other optical data. This article presents a state of the art of previous machine learning approaches, reviews the various deep learning approaches currently proposed for hyperspectral classification, and identifies the problems and difficulties which arise to implement deep neural networks for this task. In particular, the issues of spatial and spectral resolution, data volume, and transfer of models from multimedia images to hyperspectral data are addressed. Additionally, a comparative study of various families of network architectures is provided and a software toolbox is publicly released to allow experimenting with these methods. 1 This article is intended for both data scientists with interest in hyperspectral data and remote sensing experts eager to apply deep learning techniques to their own dataset.

show abstract

Section: B Spectral Classificationmentioning

confidence: 99%

Deep Learning for Classification of Hyperspectral Data: A Comparative Review

Audebert

Saux

Lefèvre

2019

IEEE Geosci. Remote Sens. Mag.

499

194

View full text Add to dashboard Cite

show abstract

“…In this section, seven different methods are used to verify the effectiveness of the proposed method, including EMAPs [8], 3D CNN [37], semi-supervised stacked autoencoders (semiSAE) [32], semi-supervised multinomial logistic regression (semiMLR) [26], maximizer of the posterior marginal by loopy belief propagation with AL (MPM-LBP-AL) [30], Mahalanobis metric for clustering (MMC) [44] and perceptual loss (PL) [47] methods.…”

Section: Comparisons Of Resultsmentioning

confidence: 99%

“…Sufficient training samples play an important role in supervised learning, however, the acquisition of available training samples in remote sensing image is difficult and time-consuming. Hence researchers focused on the semi-supervised learning, such as semi-supervised random forest [23], Transductive SVM (TSVM) [24], [25], semi-supervised multinomial logistic regression [26], ladder networks [27], self-training [28], co-training [29], active learning [30], [31], semi-supervised stacked autoencoders [32], generative adversarial networks (GAN) [33] and so on. Semisupervised learning method can improve the classification performance by combining the labeled samples and highreliable unlabeled samples.…”

Section: Introductionmentioning

confidence: 99%

Improved Metric Learning With the CNN for Very-High-Resolution Remote Sensing Image Classification

Cheng

Shen

et al. 2021

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

View full text Add to dashboard Cite

The number of labeled samples has a great impact on the classification results of very high-resolution (VHR) remote sensing image. However, the acquisition of available labeled samples is difficult and time-consuming. Faced with the limited labeled samples on high-resolution remote sensing image, semisupervised method becomes an effective way. In semi-supervised learning, accurate similarity prediction between unlabeled and labeled samples is very important. However, reliable similarity prediction between high-dimensional features is difficult. For more reliable similarity prediction for high-dimensional feature, a novel semi-supervised classification framework via improved metric learning (IML) with convolutional neural network (CNN) is proposed. In the proposed method, a novel trainable metric learning network is designed to accurately evaluate the similarity between high-dimension features. The vector distance parameter solving problem is transformed into a neural network design problem, which can automatically calculate parameters by BP algorithm. Finally, the pixel constraint mechanism is introduced to select the unlabeled samples. Experimental results conducted on three VHR remote sensing images, including Aerial, Xi'an, and Pavia University, and the results present that the proposed method performs better than the compared state-of-the-art methods.

show abstract

“…The semi-supervised classification methods include TSVM [25], semi-supervised MLR (SemiMLR) [26], a semi-supervised SAE (SemiSAE) [27], and a ladder network [28].…”

Section: Experiments Setupmentioning

confidence: 99%

“…Semi-supervised learning is an effective tool that works by combining the limited labeled samples with the highly-reliable unlabeled samples. Semi-supervised learning has been widely applied in remote sensing image classification; for instance, see the semi-supervised random forest [23], semi-supervised deep fuzzy C-mean clustering [24], transudative SVM [25], semi-supervised multinomial logistic regression (MLR) [26], semi-supervised SAEs [27], and ladder networks [28]. In [29], Zheng et al proposed a geometric low-rank Laplacian regularized semi-supervised classifier to exploit the spatial and spectral structure of HSI data.…”

Section: Introductionmentioning

confidence: 99%

Hierarchical Multi-View Semi-Supervised Learning for Very High-Resolution Remote Sensing Image Classification

Cheng

Yang

et al. 2020

Remote Sensing

View full text Add to dashboard Cite

Traditional classification methods used for very high-resolution (VHR) remote sensing images require a large number of labeled samples to obtain higher classification accuracy. Labeled samples are difficult to obtain and costly. Therefore, semi-supervised learning becomes an effective paradigm that combines the labeled and unlabeled samples for classification. In semi-supervised learning, the key issue is to enlarge the training set by selecting highly-reliable unlabeled samples. Observing the samples from multiple views is helpful to improving the accuracy of label prediction for unlabeled samples. Hence, the reasonable view partition is very important for improving the classification performance. In this paper, a hierarchical multi-view semi-supervised learning framework with CNNs (HMVSSL) is proposed for VHR remote sensing image classification. Firstly, a superpixel-based sample enlargement method is proposed to increase the number of training samples in each view. Secondly, a view partition method is designed to partition the training set into two independent views, and the partitioned subsets are characterized by being inter-distinctive and intra-compact. Finally, a collaborative classification strategy is proposed for the final classification. Experiments are conducted on three VHR remote sensing images, and the results show that the proposed method performs better than several state-of-the-art methods.

show abstract

Semi-supervised classification of hyperspectral imagery based on stacked autoencoders

Cited by 10 publications

References 7 publications

Deep Learning for Classification of Hyperspectral Data: A Comparative Review

Deep Learning for Classification of Hyperspectral Data: A Comparative Review

Improved Metric Learning With the CNN for Very-High-Resolution Remote Sensing Image Classification

Hierarchical Multi-View Semi-Supervised Learning for Very High-Resolution Remote Sensing Image Classification

Contact Info

Product

Resources

About