A novel discriminative dictionary learning method for image classification

摘要

In this paper, we present a novel discriminative dictionarylearning (DDL) method for image classification. The local structural relationshipbetween samples is first built by the Laplacian eigenmaps (LE),and then integrated into the basic DDL frame to suppress inter-class ambiguityin the feature space. Moreover, in order to improve the discriminativeability of the dictionary, the category label information of training samplesis formulated into the objective function of dictionary learning by consideringthe discriminative promotion term. Thus, the data points of originalsamples are transformed into a new feature space, in which the points fromdifferent categories are expected to be far apart. The test results based onthe real dataset indicate the effectiveness of this method.