Human-centric metrics for indoor scene assessment and synthesis

摘要: Abstract In recent years, many approaches have been proposed to analyze the affordance of 3D objects and facilitate synthesis aesthetic realistic indoor scenes. However, how assess quality such synthesized scenes is still a challenging problem. To address this, we present novel approach through so-called Human-Centric Metrics (HCMs) for quantitatively evaluating layout certain objects, thus scene synthesis. Our HCMs consider both human-object factors functional accessibility human-environment subjective comfort in scene. Given with arranged furniture our method automatically places human agents then use object arrangement environment based on position/direction agent. The conducted user study shows that assessment capability well conforms interior design knowledge. We also results synthesize by suggesting their given an empty room. A series experimental comparisons are presented validate usability effectiveness virtual