Dynamic Learning of SCRF for Feature Selection and Classification of Hyperspectral Imagery
This paper investigates the feature selection and contextual classification of hyperspectral images through the sparse conditional random field (SCRF) model. To relieve the heavy degeneration of classification performance caused by the characteristics of the hyperspectral data and the oversparsity when SCRF selects a small feature subset, we develop a dynamic learning framework to train the SCRF. Under the piecewise training framework, the proposed dynamic learning method of SCRF can be implemented efficiently through separated dynamic sparse trainings of simple classifiers defined by corresponding potentials. Experiments on the real-world hyperspectral images attest to the effectiveness of the proposed method.
KeywordsConditional random field classification feature selection
- 7.Lafferty, J., McCallum, A., Pereira, F.: Conditional random fields: probabilistic models for segmenting and labeling sequence data. In: International Conference on Machine Learning, pp. 282–289 (2001)Google Scholar
- 10.Kumar, S.: Models for learning spatial interactions in natural images for context-based classification. PhD thesis. Carnegie Mellon University (2005)Google Scholar
- 12.Ng, A.Y.: Feature selection, L1 vs. L2 regularization, and rotational invariance. In: International Conference on Machine Learning (2004)Google Scholar