3D Human-Human Interaction Anomaly Detection

2025-12-15

AI Summary

Existing single-person anomaly detection models struggle to capture complex, asymmetric dynamic relationships in human–human interactions, leading to poor anomaly identification accuracy in collaborative scenarios. To address this limitation, we propose the novel task of 3D Human–Human Interaction Anomaly Detection (H2IAD), the first to move beyond the single-person paradigm. Our method introduces a Temporal Attention Sharing Module (TASM) to jointly model synchronized motion patterns of two individuals and a Distance-based Relation Encoding Module (DREM) to explicitly encode socially grounded spatial constraints governing their interaction. Anomaly scores are computed via normalizing flows. Evaluated on a multi-person interaction action benchmark, our approach significantly outperforms single-person baselines, demonstrating the effectiveness of high-fidelity modeling of interactive dynamics. This work establishes a foundation for anomaly detection in socially situated, cooperative human activities.

Technology Category

• Humans and AI: Interaction Techniques and Devices
• Cognitive Modeling & Cognitive Systems: Social Cognition And Interaction
• Intelligent Robots: Human-Robot Interaction

Application Category

• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Graph Algorithms and Modeling for the Web: Algorithms and analysis for heterogeneous, signed, attributed, multi-relational, temporal, higher-order, and annotated Web-related graphs
• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets

Abstract

Human-centric anomaly detection (AD) has been primarily studied to specify anomalous behaviors in a single person. However, as humans by nature tend to act in a collaborative manner, behavioral anomalies can also arise from human-human interactions. Detecting such anomalies using existing single-person AD models is prone to low accuracy, as these approaches are typically not designed to capture the complex and asymmetric dynamics of interactions. In this paper, we introduce a novel task, Human-Human Interaction Anomaly Detection (H2IAD), which aims to identify anomalous interactive behaviors within collaborative 3D human actions. To address H2IAD, we then propose Interaction Anomaly Detection Network (IADNet), which is formalized with a Temporal Attention Sharing Module (TASM). Specifically, in designing TASM, we share the encoded motion embeddings across both people such that collaborative motion correlations can be effectively synchronized. Moreover, we notice that in addition to temporal dynamics, human interactions are also characterized by spatial configurations between two people. We thus introduce a Distance-Based Relational Encoding Module (DREM) to better reflect social cues in H2IAD. The normalizing flow is eventually employed for anomaly scoring. Extensive experiments on human-human motion benchmarks demonstrate that IADNet outperforms existing Human-centric AD baselines in H2IAD.

Problem

• Detects anomalies in collaborative 3D human-human interactions
• Addresses limitations of single-person anomaly detection models
• Captures complex temporal and spatial dynamics in interactions

Innovation

• Temporal Attention Sharing Module for motion synchronization
• Distance-Based Relational Encoding for spatial social cues
• Normalizing flow for anomaly scoring in interactions