ReMoS: 3D Motion-Conditioned Reaction Synthesis for Two-Person Interactions

arXiv:2311.17057v2 Announce Type: replace
Abstract: Current approaches for 3D human motion synthesis generate high-quality animations of digital humans performing a wide variety of actions and gestures. However, a notable technological gap exists in addressing the complex dynamics of multi-human interactions within this paradigm. In this work, we present ReMoS, a denoising diffusion-based model that synthesizes full-body reactive motion of a person in a two-person interaction scenario. Assuming the motion of one person is given, we employ a combined spatio-temporal cross-attention …

abstract animations arxiv cs.cv current denoising diffusion digital digital humans dynamics gap generate gestures however human human interactions humans interactions paradigm person quality synthesis type work