Gaze-Enhanced Multimodal Turn-Taking Prediction in Triadic Conversations

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Heo, Seongsil, Murdock, Calvin, Proulx, Michael, Miller, Christi
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910973120479232
author Heo, Seongsil
Murdock, Calvin
Proulx, Michael
Miller, Christi
author_facet Heo, Seongsil
Murdock, Calvin
Proulx, Michael
Miller, Christi
contents 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.
format Preprint
id arxiv_https___arxiv_org_abs_2505_13688
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Gaze-Enhanced Multimodal Turn-Taking Prediction in Triadic Conversations
Heo, Seongsil
Murdock, Calvin
Proulx, Michael
Miller, Christi
Human-Computer Interaction
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.
title Gaze-Enhanced Multimodal Turn-Taking Prediction in Triadic Conversations
topic Human-Computer Interaction
url https://arxiv.org/abs/2505.13688