Face-to-Face: A Video Dataset for Multi-Person Interaction Modeling

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Chu, Ernie, Patel, Vishal M.
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911555256320000
author Chu, Ernie
Patel, Vishal M.
author_facet Chu, Ernie
Patel, Vishal M.
contents Modeling the reactive tempo of human conversation remains difficult because most audio-visual datasets portray isolated speakers delivering short monologues. We introduce \textbf{Face-to-Face with Jimmy Fallon (F2F-JF)}, a 70-hour, 14k-clip dataset of two-person talk-show exchanges that preserves the sequential dependency between a guest turn and the host's response. A semi-automatic pipeline combines multi-person tracking, speech diarization, and lightweight human verification to extract temporally aligned host/guest tracks with tight crops and metadata that are ready for downstream modeling. We showcase the dataset with a reactive, speech-driven digital avatar task in which the host video during $[t_1,t_2]$ is generated from their audio plus the guest's preceding video during $[t_0,t_1]$. Conditioning a MultiTalk-style diffusion model on this cross-person visual context yields small but consistent Emotion-FID and FVD gains while preserving lip-sync quality relative to an audio-only baseline. The dataset, preprocessing recipe, and baseline together provide an end-to-end blueprint for studying dyadic, sequential behavior, which we expand upon throughout the paper. Dataset and code are available at https://face2face2026.github.io.
format Preprint
id arxiv_https___arxiv_org_abs_2603_14794
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Face-to-Face: A Video Dataset for Multi-Person Interaction Modeling
Chu, Ernie
Patel, Vishal M.
Computer Vision and Pattern Recognition
Machine Learning
Modeling the reactive tempo of human conversation remains difficult because most audio-visual datasets portray isolated speakers delivering short monologues. We introduce \textbf{Face-to-Face with Jimmy Fallon (F2F-JF)}, a 70-hour, 14k-clip dataset of two-person talk-show exchanges that preserves the sequential dependency between a guest turn and the host's response. A semi-automatic pipeline combines multi-person tracking, speech diarization, and lightweight human verification to extract temporally aligned host/guest tracks with tight crops and metadata that are ready for downstream modeling. We showcase the dataset with a reactive, speech-driven digital avatar task in which the host video during $[t_1,t_2]$ is generated from their audio plus the guest's preceding video during $[t_0,t_1]$. Conditioning a MultiTalk-style diffusion model on this cross-person visual context yields small but consistent Emotion-FID and FVD gains while preserving lip-sync quality relative to an audio-only baseline. The dataset, preprocessing recipe, and baseline together provide an end-to-end blueprint for studying dyadic, sequential behavior, which we expand upon throughout the paper. Dataset and code are available at https://face2face2026.github.io.
title Face-to-Face: A Video Dataset for Multi-Person Interaction Modeling
topic Computer Vision and Pattern Recognition
Machine Learning
url https://arxiv.org/abs/2603.14794