Semisupervised classification of hyperspectral images based on tri-training algorithm with enhanced diversity

Abstract. Hyperspectral image classification faces a serious challenge due to the high dimension of hyperspectral data and limited labeled samples. Tri-training algorithm is a widely used semisupervised classification method, but the algorithm lacks significant diversity among the classifiers when the number of initial label samples is limited. A semisupervised classification method for hyperspectral data based on tri-training is proposed. It combines different classifiers and stratified sampling based on labeled class to increase classifier diversity and avoid classifier performance deterioration. Performance comparison between the proposed algorithm and tri-training algorithm was made through experiments. The proposed algorithm improved the overall accuracy and Kappa coefficient by 1.37% to 6.84% and 0.0096 to 0.0808, respectively, and the results showed that the effectiveness of the algorithm is verified. Moreover, the algorithm can also get better performance when the number of samples is small.

[1]  Lorenzo Bruzzone,et al.  A multiscale expectation-maximization semisupervised classifier suitable for badly posed image classification , 2006, IEEE Transactions on Image Processing.

[2]  Qian Du,et al.  An efficient semi-supervised classification approach for hyperspectral imagery , 2014 .

[3]  Daniel Jiwoong Im,et al.  Semisupervised Hyperspectral Image Classification via Neighborhood Graph Learning , 2015, IEEE Geoscience and Remote Sensing Letters.

[4]  Rui Huang,et al.  Using tri-training to exploit spectral and spatial information for hyperspectral data classification , 2012, 2012 International Conference on Computer Vision in Remote Sensing.

[5]  Rui Zhang,et al.  Semi-Supervised Hyperspectral Image Classification Using Spatio-Spectral Laplacian Support Vector Machine , 2014, IEEE Geoscience and Remote Sensing Letters.

[6]  Wang Jingyan.  A Semi-supervised Collaboration Classification Algorithm with Enhanced Difference , 2013 .

[7]  Yan Zhou,et al.  Enhancing Supervised Learning with Unlabeled Data , 2000, ICML.

[8]  Shaohui Mei,et al.  Hyperspectral image classification based on iterative Support Vector Machine by integrating spatial-spectral information , 2013, 2013 IEEE International Geoscience and Remote Sensing Symposium - IGARSS.

[9]  颜露新,et al.  Spatial–spectral method for classification of hyperspectral images , 2013 .

[10]  Dae-Ki Kang,et al.  Detecting the spam review using tri-training , 2015, 2015 17th International Conference on Advanced Communication Technology (ICACT).

[11]  Qian Du,et al.  Robust Hyperspectral Image Classification by Multi-Layer Spatial-Spectral Sparse Representations , 2016, Remote. Sens..

[12]  Avrim Blum,et al.  The Bottleneck , 2021, Monopsony Capitalism.

[13]  Zhi-Hua Zhou,et al.  Tri-training: exploiting unlabeled data using three classifiers , 2005, IEEE Transactions on Knowledge and Data Engineering.

[14]  Hai Jin,et al.  MSVM-kNN: Combining SVM and k-NN for Multi-class Text Classification , 2008, IEEE International Workshop on Semantic Computing and Systems.

[15]  Li Guo Wang,et al.  Semi-Supervised Classification for Hyperspectral Image Based on Tri-Training , 2014, CIT 2014.