Mutually exclusive-KSVD: Learning a discriminative dictionary for hyperspectral image classification

Sparse representation and dictionary learning methods have been successfully applied in classification of hyperspectral images (HSIs). However, when the number of training data is insufficient which is widely happened in HSI classification, the learned sparse representation is generally insufficient and the corresponding performances would be significantly degraded. To address the above problem, in this paper, we propose a novel dictionary learning method, namely mutually exclusive K-SVD. We construct a mutual exclusion term for the dictionary by decomposing each class of sub-dictionary into positive and negative categories. Therefore, the learned sparse codes not only consider the within-class consistency, but also between-class mutual exclusion, thereby resulting in improved classification performance with limited training samples. Furthermore, in the testing phase, we utilize the multiscale strategy for each pixel instead of pixel-wise coding to make full use of the spatial features of the image and further improve the classification accuracy. Experimental results demonstrate that the proposed algorithm outperforms state-of-the art algorithms in both qualitative and quantitative evaluations.