Perceiving and Acting in First-Person: A Dataset and Benchmark for Egocentric Human-Object-Human Interactions

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Xu, Liang, Yang, Chengqun, Lin, Zili, Xu, Fei, Liu, Yifan, Xu, Congsheng, Zhang, Yiyi, Qin, Jie, Sheng, Xingdong, Liu, Yunhui, Jin, Xin, Yan, Yichao, Zeng, Wenjun, Yang, Xiaokang
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915432494006272
author Xu, Liang
Yang, Chengqun
Lin, Zili
Xu, Fei
Liu, Yifan
Xu, Congsheng
Zhang, Yiyi
Qin, Jie
Sheng, Xingdong
Liu, Yunhui
Jin, Xin
Yan, Yichao
Zeng, Wenjun
Yang, Xiaokang
author_facet Xu, Liang
Yang, Chengqun
Lin, Zili
Xu, Fei
Liu, Yifan
Xu, Congsheng
Zhang, Yiyi
Qin, Jie
Sheng, Xingdong
Liu, Yunhui
Jin, Xin
Yan, Yichao
Zeng, Wenjun
Yang, Xiaokang
contents Learning action models from real-world human-centric interaction datasets is important towards building general-purpose intelligent assistants with efficiency. However, most existing datasets only offer specialist interaction category and ignore that AI assistants perceive and act based on first-person acquisition. We urge that both the generalist interaction knowledge and egocentric modality are indispensable. In this paper, we embed the manual-assisted task into a vision-language-action framework, where the assistant provides services to the instructor following egocentric vision and commands. With our hybrid RGB-MoCap system, pairs of assistants and instructors engage with multiple objects and the scene following GPT-generated scripts. Under this setting, we accomplish InterVLA, the first large-scale human-object-human interaction dataset with 11.4 hours and 1.2M frames of multimodal data, spanning 2 egocentric and 5 exocentric videos, accurate human/object motions and verbal commands. Furthermore, we establish novel benchmarks on egocentric human motion estimation, interaction synthesis, and interaction prediction with comprehensive analysis. We believe that our InterVLA testbed and the benchmarks will foster future works on building AI agents in the physical world.
format Preprint
id arxiv_https___arxiv_org_abs_2508_04681
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Perceiving and Acting in First-Person: A Dataset and Benchmark for Egocentric Human-Object-Human Interactions
Xu, Liang
Yang, Chengqun
Lin, Zili
Xu, Fei
Liu, Yifan
Xu, Congsheng
Zhang, Yiyi
Qin, Jie
Sheng, Xingdong
Liu, Yunhui
Jin, Xin
Yan, Yichao
Zeng, Wenjun
Yang, Xiaokang
Computer Vision and Pattern Recognition
Learning action models from real-world human-centric interaction datasets is important towards building general-purpose intelligent assistants with efficiency. However, most existing datasets only offer specialist interaction category and ignore that AI assistants perceive and act based on first-person acquisition. We urge that both the generalist interaction knowledge and egocentric modality are indispensable. In this paper, we embed the manual-assisted task into a vision-language-action framework, where the assistant provides services to the instructor following egocentric vision and commands. With our hybrid RGB-MoCap system, pairs of assistants and instructors engage with multiple objects and the scene following GPT-generated scripts. Under this setting, we accomplish InterVLA, the first large-scale human-object-human interaction dataset with 11.4 hours and 1.2M frames of multimodal data, spanning 2 egocentric and 5 exocentric videos, accurate human/object motions and verbal commands. Furthermore, we establish novel benchmarks on egocentric human motion estimation, interaction synthesis, and interaction prediction with comprehensive analysis. We believe that our InterVLA testbed and the benchmarks will foster future works on building AI agents in the physical world.
title Perceiving and Acting in First-Person: A Dataset and Benchmark for Egocentric Human-Object-Human Interactions
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2508.04681