Scene segmentation using texture, color, motion; representation of 2-D or 3-D structures; knowledge-based vision systems.