Color Texture Histograms for Natural Images Interpretation

This paper presents a recognition method for natural images based on color texture histograms inthe context of image interpretation and scene modeling. A color histogram of sums and differencesis proposed to obtain texture features which are faster to compute than correlograms ( i.e., coloredversion of co-occurrence matrices) and improving substantially object recognition. Outdoor naturalimages are generally affected by color casting artifacts which can affect object recognition. Therefore, an on-line color balancing algorithm based on chromatic adaptation models, eliminatesthese color deviations. The proposed approach globally involves functions as color segmentation,histogram texture analysis and a region recognition step. Our approach has been extensively testedand validated to obtain an accurate 2D scene interpretation from natural images. This techniquemay be used in robot navigation by identifying navigable regions ( e.g., roads or fairly flat surfaces) on natural scenes, scene modeling and image categorization.