Synchrony Vision: An RGB-D Sensor-Based System for Real-Time Monitoring and Event-Level Analysis of Interpersonal...

Area of Science:

• Human-Computer Interaction
• Biomedical Engineering
• Psychology

Background:

• Interpersonal synchrony involves time-dependent coordination of interacting partners' body movements.
• Quantifying this synchrony in naturalistic settings remains challenging.

Purpose of the Study:

• To present Synchrony Vision, an RGB-D sensor-based system for real-time monitoring and event-level analysis of interpersonal motion synchrony.
• To evaluate the system's effectiveness in analyzing free-dialog conversations.

Main Methods:

• The system transforms RGB-D sensor skeletal data into joint acceleration signals.
• It estimates event-level phase differences and synchrony metrics within a 1.0 s window.
• A deployed pipeline was evaluated on 25 Kinect-tracked dyads during unconstrained conversation.

Main Results:

• Synchrony Vision detected 2681 synchrony events across 200.6 minutes of conversation.
• The observed event rate significantly exceeded circular-surrogate baselines.
• Dyad rankings for synchrony were stable across parameter settings, though different from Motion Energy Analysis.

Conclusions:

• RGB-D skeletal sensing enables transparent, event-level quantification of interpersonal coordination.
• This technology extends human motion analysis beyond individual movement capture.
• Synchrony Vision provides a novel tool for studying social interaction dynamics.