A deep feature manifold embedding method for hyperspectral image classification

ABSTRACT In this letter, we proposed a novel deep feature manifold embedding method to improve feature extraction ability of traditional deep learning methods. This method first obtains deep features of hyperspectral image (HSI) from a trained autoencoder. Then, an intrinsic graph and a penalty graph are constructed to discover the discriminant manifold structure of deep features. Finally, the deep features are mapped into a low-dimensional embedding space, in which samples in intraclass manifold are compacted and samples from interclass manifolds are separated. Experiments on Pavia University, Indian Pines and Urban datasets demonstrate that the proposed method effectively improves the classification performance of HSI compared with other state-of-the-art approaches.

[1]  Xin Zhang,et al.  Image recognition method based on supervised multi-manifold learning , 2017, J. Intell. Fuzzy Syst..

[2]  R. Fisher.  THE USE OF MULTIPLE MEASUREMENTS IN TAXONOMIC PROBLEMS , 1936 .

[3]  Xing Zhao,et al.  Spectral–Spatial Classification of Hyperspectral Data Based on Deep Belief Network , 2015, IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing.

[4]  Gang Yang,et al.  A Sparse and Low-Rank Near-Isometric Linear Embedding Method for Feature Extraction in Hyperspectral Imagery Classification , 2017, IEEE Transactions on Geoscience and Remote Sensing.

[5]  Licheng Jiao,et al.  Recursive Autoencoders-Based Unsupervised Feature Learning for Hyperspectral Image Classification , 2017, IEEE Geoscience and Remote Sensing Letters.

[6]  Richard J. Murphy,et al.  Unsupervised Feature-Learning for Hyperspectral Data with Autoencoders , 2019, Remote. Sens..

[7]  Karl Pearson F.R.S..  LIII. On lines and planes of closest fit to systems of points in space , 1901 .

[8]  Xiaofei He,et al.  Locality Preserving Projections , 2003, NIPS.

[9]  Gang Wang,et al.  Deep Learning-Based Classification of Hyperspectral Data , 2014, IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing.

[10]  Jiamin Liu,et al.  Local Geometric Structure Feature for Dimensionality Reduction of Hyperspectral Imagery , 2017, Remote. Sens..

[11]  XLIX. A new specifying method for stress and strain in an elastic solid , 1902 .

[12]  Bo Du,et al.  Feature Learning Using Spatial-Spectral Hypergraph Discriminant Analysis for Hyperspectral Image , 2019, IEEE Transactions on Cybernetics.

[13]  W. Wong,et al.  Supervised optimal locality preserving projection , 2012, Pattern Recognit..

[14]  Stephen Lin,et al.  Graph Embedding and Extensions: A General Framework for Dimensionality Reduction , 2007, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[15]  Guigang Zhang,et al.  Deep Learning , 2016, Int. J. Semantic Comput..

[16]  Mengdi Wang,et al.  Unsupervised feature extraction for hyperspectral images using combined low rank representation and locally linear embedding , 2017, 2017 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP).

[17]  Hong Huang,et al.  Improved discriminant sparsity neighborhood preserving embedding for hyperspectral image classification , 2014, Neurocomputing.

[18]  Peng Zhang,et al.  A nonlinear and explicit framework of supervised manifold-feature extraction for hyperspectral image classification , 2019, Neurocomputing.

[19]  Stephen Marshall,et al.  Segmented Autoencoders for Unsupervised Embedded Hyperspectral Band Selection , 2018, 2018 7th European Workshop on Visual Information Processing (EUVIP).

[20]  Geoffrey E. Hinton,et al.  Deep Learning , 2015, Nature.