Color image segmentation by supervised pixel classification in a color texture feature space. Application to soccer image segmentation

We describe an approach to color image segmentation which is considered as a supervised pixel classification problem. The pixel classification algorithm analyses the color texture features, that is to say the texture features which are computed by tacking into account the color components of the neighbor pixels. We determine the most discriminating color texture features among a multidimensional set of color texture features by means of an iterative feature selection procedure associated to an information criterion. We successfully apply our approach to soccer image segmentation