Multi-Person Interaction Generation from Two-Person Motion Priors

2025-05-23

AI Summary

Existing methods for generating multi-person social interactions require retraining models per scenario and suffer from interpenetration artifacts and motion mode collapse. Method: This paper proposes a graph-driven interactive sampling framework that decomposes multi-person motion into a Pairwise Interaction Graph (PIG) composed of dyadic interactions, leveraging a pretrained two-person motion diffusion model as a prior—eliminating the need for dedicated multi-person model training. Contribution/Results: We introduce novel graph-structured modeling and a dual-conditioning guidance mechanism that jointly enforces kinematic feasibility and graph-structured interaction dependencies during sampling. This effectively suppresses interpenetration and enhances motion diversity. Experiments demonstrate significant improvements over state-of-the-art methods across diverse two-person and multi-person interaction scenarios, enabling high-fidelity, controllable, and diverse multi-agent coordinated motion synthesis.

Technology Category

• Humans and AI: Interaction Techniques and Devices
• Cognitive Modeling & Cognitive Systems: Social Cognition And Interaction
• Computer Vision: Diffusion Models for Vision

Application Category

• User Modeling, Personalization and Recommendation: User modeling and simulation for interactive and conversational systems
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Generating realistic human motion with high-level controls is a crucial task for social understanding, robotics, and animation. With high-quality MOCAP data becoming more available recently, a wide range of data-driven approaches have been presented. However, modelling multi-person interactions still remains a less explored area. In this paper, we present Graph-driven Interaction Sampling, a method that can generate realistic and diverse multi-person interactions by leveraging existing two-person motion diffusion models as motion priors. Instead of training a new model specific to multi-person interaction synthesis, our key insight is to spatially and temporally separate complex multi-person interactions into a graph structure of two-person interactions, which we name the Pairwise Interaction Graph. We thus decompose the generation task into simultaneous single-person motion generation conditioned on one other's motion. In addition, to reduce artifacts such as interpenetrations of body parts in generated multi-person interactions, we introduce two graph-dependent guidance terms into the diffusion sampling scheme. Unlike previous work, our method can produce various high-quality multi-person interactions without having repetitive individual motions. Extensive experiments demonstrate that our approach consistently outperforms existing methods in reducing artifacts when generating a wide range of two-person and multi-person interactions.

Problem

• Generating realistic multi-person interactions from two-person motion priors
• Reducing artifacts like body interpenetrations in generated interactions
• Decomposing complex interactions into pairwise graphs for diverse outputs

Innovation

• Leverages two-person motion diffusion models
• Uses Pairwise Interaction Graph structure
• Introduces graph-dependent guidance terms