Ego4D

Ego4D is a massive-scale egocentric video dataset and benchmark suite introduced by Meta AI in 2021. It offers 3,670 hours of daily-life activity video spanning hundreds of scenarios (household, outdoor, workplace, leisure, social gatherings, etc.) captured by 931 unique camera wearers from 74 worldwide locations across 9 countries. The dataset is organized around five benchmark challenges: episodic memory, hand-object interaction forecasting, audio-visual diarization, social interaction understanding, and manipulation skill proficiency. The collection process was designed with rigorous privacy and ethics standards, including consenting participants and robust de-identification procedures. Ego4D has become a landmark dataset in egocentric vision, embodied AI, and human activity understanding, with applications spanning robot learning, augmented reality, and assistive technology. Its scale and diversity dramatically expands the publicly available egocentric video data for the research community.