Shape up! Perception based body shape variation for data-driven crowds

摘要: Representative distribution of body shapes is needed when simulating crowds in real-world situations, e.g., for city or event planning. Visual realism and plausibility are often also required visualization purposes, while these the top criteria entertainment applications such as games movie production. Therefore, achieving representative visually plausible body-shape variation optimizing available resources an important goal. We present a data-driven approach to generating selecting models with varied shapes, based on measurement demographic data from CAESAR anthropometric database. conducted online perceptual study explore relationship between shape, distinctiveness attractiveness bodies close median height girth. found that most salient differences size upper-lower ratios, particular respect shoulders, waist hips. Based results, we propose strategies shape selection have validated lab-based study. Finally, demonstrate our results crowd system perceptually distribution.