Hyperspectral Image Clustering Based on Unsupervised Broad Learning

Yi, Kyu Yang; Cheng, Yuhu; Chen, C. L. Philip; Wang, Xuesong

doi:10.1109/lgrs.2019.2907598

Cited by 36 publications

(18 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Kong et al [40] proposed a semisupervised model by merging the class-probability structure into BLS and achieved good classification performance in HSI classification. Kong et al [41] proposed a HSI clustering algorithm based on BLS, and exploited the graph-regularized sparse autoencoder to fine-tune the weights of MF and EN.…”

Section: Introductionmentioning

confidence: 99%

Semisupervised Classification of Hyperspectral Image Based on Graph Convolutional Broad Network

Wang

Cheng

Chen

et al. 2021

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

Self Cite

View full text Add to dashboard Cite

Hyperspectral image (HSI) classification has attracted much attention in the field of remote sensing. However, the lack of sufficient labeled training samples is a huge challenge for HSI classification. To face this challenge, we propose a semi-supervised HSI classification method based on graph convolutional broad network (GCBN). Firstly, to avoid the underfitting problem caused by the insufficient linear sparse feature representation ability of broad learning system (BLS), graph convolution operation is applied to extract non-linear and discriminative spectral-spatial features from the original HSI to replace the linear mapping features in the traditional BLS. Secondly, to solve the problem of insufficient model classification ability caused by limited labeled samples, the combinatorial average method (CAM) is proposed to use valuable paired samples to generate sample expansion set for GCBN model training. Thirdly, BLS is used to perform broad expansion on spectral-spatial features extracted by GCN and extended by CAM, which further enhances the feature representation ability. Finally, the output weights can be easily calculated by the ridge regression theory. Experimental results on three real HSI datasets demonstrate the effectiveness of our proposed GCBN.

show abstract

Section: Introductionmentioning

confidence: 99%

Semisupervised Classification of Hyperspectral Image Based on Graph Convolutional Broad Network

Wang

Cheng

Chen

et al. 2021

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

Self Cite

View full text Add to dashboard Cite

show abstract

“…Inspired by the success of deep clustering, the pioneer work of deep HSI clustering method, namely LSSD, is proposed in [1] and the discriminative deep embedding of HSI samples are learned by training a siamese network based on the (dis)similar sample pairs derived from the set-to-set and sample-to-sample distances. The unsupervised broad learning is first applied for HSI clustering by using a graph regularized sparse autoencoder [35].…”

Section: Introductionmentioning

confidence: 99%

Affinity Matrix Learning Via Nonnegative Matrix Factorization for Hyperspectral Imagery Clustering

Qin

et al. 2021

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

View full text Add to dashboard Cite

In this paper, we integrate the spatial-spectral information of HSI samples into non-negative matrix factorization (NMF) for affinity matrix learning to address the issue of HSI clustering. This technique consists of three main components: i) oversegmentation for computing the spectral-spatial affinity matrix, ii) NMF with the guidance of the obtained affinity matrix and iii) density-based spectral clustering on the final affinity matrix. First, the HSI is oversegmented into superpixels via the entropy rate superpixel (ERS) algorithm. The spectral-spatial affinity matrix is defined based on the class-consistency assumption of all the HSI samples in each superpixel and the similar HSI samples between adjacent superpixels. Second, to integrate the spectral-spatial information into NMF, the obtained affinity matrix is used to guide the iterative process of NMF. The spectralspatial affinity matrix is then weighted by the affinity matrix in the obtained low-dimensional subspace to form the final affinity matrix. Third, density-based spectral clustering is applied to the final affinity matrix to obtain clustering maps. Experimental results on three public benchmark HSIs demonstrate that the proposed method is superior to the considered state-of-the-art baseline methods on both the computational cost and clustering accuracy.

show abstract

“…Kong et al [33] merged the class-probability structure into BLS to obtain a semi-supervised learning version, which achieves a high accuracy in HSI classification. Kong et al [34] fine-tuned the weights of MF and EN with the graphregularized sparse autoencoder, which maintained the manifold structure of original data. However, both the above two HSI classification methods cannot help to improve the HSI classification accuracy by utilizing vast quantities of labeled samples in related domains.…”

Section: Introductionmentioning

confidence: 99%

Hyperspectral Image Classification Based on Domain Adaptation Broad Learning

Wang

Chen

et al. 2020

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

Self Cite

View full text Add to dashboard Cite

Hyperspectral images (HSI) are widely applied in numerous fields for their rich spatial and spectral information. However, in these applications, we always face the situation that the available labeled samples are limited or absent. Therefore, we propose a HSI classification method based on domain adaptation broad learning (DABL). Firstly, according to the importance of the marginal and conditional distributions, the maximum mean discrepancy is used in mapped features to adapt these distributions between source and target domains. Meanwhile the manifold regularization is added to maintain the manifold structure of the input HSI data. Secondly, to further reduce the distribution difference and maintain manifold structure, the domain adaptation and manifold regularization are added to the output layer of DABL. Finally, the output weights can be easily calculated by the ridge regression theory. Experimental results on three real HSI datasets demonstrate the effectiveness of our proposed DABL.

show abstract

Hyperspectral Image Clustering Based on Unsupervised Broad Learning

Cited by 36 publications

References 13 publications

Semisupervised Classification of Hyperspectral Image Based on Graph Convolutional Broad Network

Semisupervised Classification of Hyperspectral Image Based on Graph Convolutional Broad Network

Affinity Matrix Learning Via Nonnegative Matrix Factorization for Hyperspectral Imagery Clustering

Hyperspectral Image Classification Based on Domain Adaptation Broad Learning

Contact Info

Product

Resources

About