Object-Shot Enhanced Grounding Network for Egocentric Video

TL;DR

This paper introduces OSGNet, a novel egocentric video grounding model that leverages object information and shot movement analysis to improve modality alignment and achieve state-of-the-art results.

Contribution

The paper proposes a new method that incorporates object extraction and shot movement analysis to enhance egocentric video grounding performance.

Findings

Achieves state-of-the-art results on three datasets.

Effectively utilizes object information for better grounding.

Leverages shot movement features to model wearer attention.

Abstract

Egocentric video grounding is a crucial task for embodied intelligence applications, distinct from exocentric video moment localization. Existing methods primarily focus on the distributional differences between egocentric and exocentric videos but often neglect key characteristics of egocentric videos and the fine-grained information emphasized by question-type queries. To address these limitations, we propose OSGNet, an Object-Shot enhanced Grounding Network for egocentric video. Specifically, we extract object information from videos to enrich video representation, particularly for objects highlighted in the textual query but not directly captured in the video features. Additionally, we analyze the frequent shot movements inherent to egocentric videos, leveraging these features to extract the wearer's attention information, which enhances the model's ability to perform modality alignment. Experiments conducted on three datasets demonstrate that OSGNet achieves state-of-the-art performance, validating the effectiveness of our approach. Our code can be found at https://github.com/Yisen-Feng/OSGNet.