自然发生的全身背景活动数据集
收藏arXiv2015-09-21 更新2024-06-21 收录
下载链接：
http://www.dgp.toronto.edu/~dustin/ backgroundactivity/
下载链接
链接失效反馈官方服务：
资源简介：
本数据集名为‘自然发生的全身背景活动数据集’，由加拿大多伦多大学创建，旨在减少手势冲突。数据集包含52人小时的全天候背景活动数据，通过多台Kinect相机、Vicon追踪器和两个高清视频相机捕捉。创建过程涉及模拟真实生活环境，如客厅，以自然方式记录参与者的活动。数据集主要应用于研究全身交互设计，优化手势识别系统，减少误识别。
This dataset is titled 'Naturally Occurring Full-Body Background Activity Dataset', developed by the University of Toronto, Canada, with the aim of reducing gesture conflicts. It contains 52 person-hours of all-day background activity data, captured via multiple Kinect cameras, Vicon trackers and two high-definition video cameras. The dataset was constructed by simulating real-life environments such as living rooms to record participants' activities in a natural manner. Its main applications include research on full-body interaction design, optimizing gesture recognition systems and reducing false recognition.
提供机构：
多伦多大学, 加拿大
创建时间：
2015-09-21
