EgoExoBench: A Benchmark for First- and Third-person View Video Understanding in MLLMs

TL;DR

EgoExoBench is a new benchmark designed to evaluate multimodal large language models on their ability to understand and reason across first-person and third-person video perspectives, addressing a key gap in current AI capabilities.

Contribution

This paper introduces EgoExoBench, the first comprehensive benchmark for egocentric-exocentric video understanding, with datasets, tasks, and evaluation of current models' cross-view reasoning abilities.

Findings

Current models perform well on single-view tasks but struggle with cross-view semantic alignment.

Models have difficulty associating viewpoints accurately across egocentric and exocentric videos.

Temporal reasoning in cross-view contexts remains a significant challenge for existing models.

Abstract

Transferring and integrating knowledge across first-person (egocentric) and third-person (exocentric) viewpoints is intrinsic to human intelligence, enabling humans to learn from others and convey insights from their own experiences. Despite rapid progress in multimodal large language models (MLLMs), their ability to perform such cross-view reasoning remains unexplored. To address this, we introduce EgoExoBench, the first benchmark for egocentric-exocentric video understanding and reasoning. Built from publicly available datasets, EgoExoBench comprises over 7,300 question-answer pairs spanning eleven sub-tasks organized into three core challenges: semantic alignment, viewpoint association, and temporal reasoning. We evaluate 13 state-of-the-art MLLMs and find that while these models excel on single-view tasks, they struggle to align semantics across perspectives, accurately associate views, and infer temporal dynamics in the ego-exo context. We hope EgoExoBench can serve as a valuable resource for research on embodied agents and intelligent assistants seeking human-like cross-view intelligence.