Color Harmonies and Contrasts Search in Art Image Collections

In this paper a novel and more complete classification of color harmonies and contrasts are presented. It is done on the base of the three main characteristics of the color, which are most closed to the human perception - hue, saturation and lightness. Functions for automatic features extraction from digital images, using MPEG-7 dominant color descriptor are defined. The results of these functions are new high-level features vectors, corresponding to outline color combinations. An experimental software system was developed and experiments were conducted. The system can be used by designers and art students for searching images, having specific harmonies and contrasts in image collections. In future we will use the system as a Web 2.0 service, which could be included in a virtual learning environment.