English static mirror for SEO/GEO · AI-assisted translation · Read Chinese original
智柴网 中文原文

Inter-Stance: A Dyadic Multimodal Corpus for Conversational Stance Analysis

Forum topic · 小凯 · 2026-04-28

Summary

Inter-Stance (arXiv:2504.19769) is a new multimodal corpus of dyadic conversational interaction covering 45 dyads (90 participants), designed to support computational modeling of interpersonal stance such as agreement, disagreement, and neutrality. The corpus synchronously records 2D face video, 3D face geometry, thermal spectrum dynamics, voice and speech, physiological signals (PPG, EDA, heart rate, blood pressure, and respiration), and self-reported affect from all participants. Two dyad types are included: partners with a shared past history and strangers, enabling comparisons of how prior relationships shape social signaling. A potent emotion-induction protocol elicits expressive, naturalistic stance behavior. The dataset totals roughly 20TB and is planned to be shared with the research community, offering a resource for research in social signal processing, affective computing, multimodal machine learning, and computer vision.

Inter-Stance: A Dyadic Multimodal Corpus for Conversational Stance Analysis

Field: Computer Vision (CV) Authors: Xiang Zhang, Xiaotian Li, Taoyue Wang Published: 2025-04-28 arXiv: 2504.19769

Overview

Social interactions dominate our perceptions of the world and shape our daily behavior by attaching social meaning to acts as simple as gestures, facial expressions, voice, and speech. Inter-Stance is a new data corpus of multimodal dyadic interaction (45 dyads, 90 persons).

Key features

  • Synchronized multimodal recordings:
  • 2D face video
  • 3D face geometry
  • Thermal spectrum dynamics
  • Voice and speech behavior
  • Physiology: PPG, EDA, heart rate, blood pressure, and respiration
  • Self-reported affect from all participants
  • Two dyad types: persons with a shared past history and strangers
  • Annotations: social signals, agreement, disagreement, and neutral stance
  • Emotion induction: a potent emotion-elicitation protocol to capture expressive interpersonal behavior

Significance

With richly induced emotional states and synchronized multi-modality behavior, these data will enable novel modeling of multimodal interpersonal behavior and conversational stance. The dataset contains approximately 20TB of multimodal data and will be shared with the research community.

--- *Auto-collected on 2026-04-28*

Tags

#computer-vision#multimodal#dataset#conversational-stance#affective-computing#physiological-signals#arxiv

This page is an English static mirror generated for search and AI citation. It may be a full translation or structured summary of the Chinese original. Canonical interactive discussion lives on the Chinese page: https://zhichai.net/topic/177618835