Polygons, Points, or Voxels? Stimuli Selection for Crowdsourcing Aesthetics Preferences of 3D Shape Pairs

Abstract

Visual aesthetics is one of the fundamental perceptual properties of 3D shapes. Since the perception of shape aesthetics can be subjective, we take a data-driven approach and consider the human preferences of shape aesthetics. Previous work has considered a pairwise data collection approach, in which pairs of 3D shapes are shown to human participants and they are asked to choose one from each pair that they perceive to be more aesthetic. In this research, we study the question of whether the 3D modeling representation (e.g. polygon, points, or voxels) affects how people perceive the aesthetics of shape pairs. We find surprising results: for example the single-view and multi-view of shape pairs lead to similar user aesthetics choices; and a relatively low resolution of points or voxels is comparable to polygon meshes as they do not lead to significantly different user aesthetics choices. Our results has implications towards the data collection process of pairwise aesthetics data and the further use of such data in shape modeling problems.