Baoyu Liang, Qile Su, Shoutai Zhu, Yuchen Liang, Chao Tong
VidEvent is a large-scale dataset designed to improve AI's understanding of dynamic events in videos, providing over 23,000 annotated events from movie recaps for research and development.
Understanding events in videos is a difficult task for artificial intelligence because these events are complex and change over time. To help with this, researchers have created VidEvent, a large dataset with over 23,000 labeled events from movie recap videos. This dataset is carefully annotated to ensure high quality and includes detailed structures and relationships among events. The dataset is intended to help develop better AI models for understanding video content, and it is freely available for researchers to use.
