Tuesday 10 April 2018 12:45 - 13:30
In a number of classification problems, the features are represented by histograms. Traditionally, histograms are compared by relatively simple distance measures such as the chi-square, the Kullback-Leibler, or the Euclidian distance. This paper proposes a likelihood ratio classifier for histogram features, that is optimal in Neyman-Pearson sense. It is based on the assumption that the bin probabilities of the histograms can be modeled by a Dirichlet distribution. A simple method to estimate the Dirichlet parameters is included. Feature selection prior to classification improves the classification performance. It will be demonstrated that the proposed classifier outperforms the chi-square distance measure.
