Fig. 6

Visual result for the number of components C in the DMNB model: generative clusters vs DPMM clustering. a Generative clusters for NLPR image datasets, where green and red denote the distributions of salient and non-salient features, respectively. b DPMM clustering for NLPR image datasets, where the number of colours and shapes of the points denote the number of components C. The appropriate number of mixture components to use in the DMNB model for saliency estimation is generally unknown, and DPMM provides an attractive alternative to current methods. We find that C=24 using DPMM on the NLPR dataset