Scene Synthesis from Human Motion

Abstract

Large-scale capture of human motion with diverse, complex scenes, while immensely useful, is often considered prohibitively costly. Meanwhile, human motion alone contains rich information about the scene they reside in and interact with. For example, a sitting human suggests the existence of a chair, and their leg position further implies the chair's pose. In this paper, we propose to synthesize diverse, semantically reasonable, and physically plausible scenes based on human motion. Our framework, Scene Synthesis from HUMan MotiON (SUMMON), includes two steps. It first uses ContactFormer, our newly introduced contact predictor, to obtain temporally consistent contact labels from human motion. Based on these predictions, SUMMON then chooses interacting objects and optimizes physical plausibility losses; it further populates the scene with objects that do not interact with humans. Experimental results demonstrate that SUMMON synthesizes feasible, plausible, and diverse scenes and has the potential to generate extensive human-scene interaction data for the community. * and † indicate equal contribution. https://sites.google.com/stanford.edu/summon

Introduction to materials chemistry at Xi’an Jiaotong University

“…Human-Scene Interaction Representation: Human-scene interaction has also been widely studied in other tasks such as human or/and scene reconstruction [14,15] and scene generation [16,17]. Dang et al [14] propose to estimate the possible region the human body can be positioned from the scene information, and use it to constrain the human pose for human pose reconstruction under severe occlusions.…”

“…Savva et al [16] propose to learn a joint distribution of human poses and object arrangements and generate a plausible interaction by sampling from the distribution. Ye et al [17] propose to generate the scene layout from the human motions. We focus on generating diverse human poses in scenes.…”