Gaze-Enhanced Multimodal Turn-Taking Prediction in Triadic Conversations

TL;DR

This paper presents a lightweight, privacy-conscious framework that integrates gaze and speaker localization to improve turn-taking prediction in triadic conversations, enhancing speech intelligibility in noisy environments.

Contribution

It introduces a novel method combining gaze and spatial cues for turn-taking prediction without heavy computation, advancing multimodal interaction modeling.

Findings

Gaze data from a single user improves prediction accuracy.

Multi-user gaze data further enhances prediction performance.

The approach supports adaptive sound control in noisy environments.

Abstract

Turn-taking prediction is crucial for seamless interactions. This study introduces a novel, lightweight framework for accurate turn-taking prediction in triadic conversations without relying on computationally intensive methods. Unlike prior approaches that either disregard gaze or treat it as a passive signal, our model integrates gaze with speaker localization, structuring it within a spatial constraint to transform it into a reliable predictive cue. Leveraging egocentric behavioral cues, our experiments demonstrate that incorporating gaze data from a single-user significantly improves prediction performance, while gaze data from multiple-users further enhances it by capturing richer conversational dynamics. This study presents a lightweight and privacy-conscious approach to support adaptive, directional sound control, enhancing speech intelligibility in noisy environments, particularly for hearing assistance in smart glasses.