Sequence-Based Identification of First-Person Camera Wearers in Third-Person Views

TL;DR

This paper introduces TF2025, a new dataset with synchronized first- and third-person videos, and proposes a sequence-based method that uses motion cues and re-identification to recognize first-person camera wearers in third-person views.

Contribution

The paper presents a novel dataset TF2025 and a sequence-based identification method for first-person wearers in third-person footage, addressing a key gap in multi-camera egocentric vision research.

Findings

TF2025 dataset enables new multi-camera interaction studies.

The proposed method effectively identifies first-person wearers in third-person videos.

Results demonstrate improved accuracy over existing approaches.

Abstract

The increasing popularity of egocentric cameras has generated growing interest in studying multi-camera interactions in shared environments. Although large-scale datasets such as Ego4D and Ego-Exo4D have propelled egocentric vision research, interactions between multiple camera wearers remain underexplored-a key gap for applications like immersive learning and collaborative robotics. To bridge this, we present TF2025, an expanded dataset with synchronized first- and third-person views. In addition, we introduce a sequence-based method to identify first-person wearers in third-person footage, combining motion cues and person re-identification.