BEAR (Behaviors for Environment and Actions Recognition dataset)
收藏资源简介:
BEAR数据集是一个面向细粒度行为识别的新视频数据集,由上海交通大学创建。该数据集专注于两个主要因素定义的行为:环境和动作。它包括两个细粒度行为协议,以及多个子协议作为不同场景。该数据集旨在提供一个公平和全面的细粒度视频行为数据集,通过控制行为的环境和动作这两个决定性因素,为行为识别领域提供严格的基准和精心制作的注释。
The BEAR Dataset is a novel video dataset for fine-grained behavior recognition, created by Shanghai Jiao Tong University. This dataset focuses on behaviors defined by two main factors: environment and action. It includes two fine-grained behavior protocols, as well as multiple sub-protocols representing different scenarios. The dataset aims to provide a fair and comprehensive fine-grained video behavior dataset, and by controlling the two decisive factors of behavior—environment and action, it offers rigorous benchmarks and meticulously crafted annotations for the field of behavior recognition.




